diff --git a/AGENTS.md b/AGENTS.md
index fdefcdb3b..dc8d950ce 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -107,7 +107,7 @@ live here at `interp-engine/`; it is now its own repository,
 with its own tests, docs, lint gates and release process. Its comparison harness against
 TransformerLens, nnsight, vLLM and SGLang went with it.
 
-Each of the three apps pins an **exact version** (`interp-engine[...]==1.3.3`) in its
+Each of the three apps pins an **exact version** (`interp-engine[...]==1.6.0`) in its
 `pyproject.toml`. Exact rather than floored because the engine is first-party and moves in lockstep
 with these services: a relock for some unrelated dependency must not silently swap the code that
 serves every activation.
diff --git a/apps/graph/pyproject.toml b/apps/graph/pyproject.toml
index 2a8cba186..c49f66f98 100644
--- a/apps/graph/pyproject.toml
+++ b/apps/graph/pyproject.toml
@@ -12,7 +12,7 @@ dependencies = [
   "pydantic>=2.11.4",
   "torch>=2.11.0",
   "transformers>=4.57.1",
-  "interp-engine==1.3.3",
+  "interp-engine==1.6.0",
   "python-dotenv>=1.1.0,<2",
   "hf-transfer>=0.1.9,<0.2",
   "psutil>=7.0,<8",
diff --git a/apps/graph/uv.lock b/apps/graph/uv.lock
index e11eddda8..bdd39f610 100644
--- a/apps/graph/uv.lock
+++ b/apps/graph/uv.lock
@@ -737,7 +737,7 @@ wheels = [
 
 [[package]]
 name = "interp-engine"
-version = "1.3.3"
+version = "1.6.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "einops" },
@@ -745,9 +745,9 @@ dependencies = [
     { name = "torch" },
     { name = "transformers" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/10/18/f82b334ef3c81de79eab7423ad273ecd7a1e110f0117b74a96c92b5e9601/interp_engine-1.3.3.tar.gz", hash = "sha256:948fdc71db9119efc484619487d8224c9309e71d05687d3696612d5cc65f3b26", size = 777436, upload-time = "2026-08-21T00:02:39.338Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/42/1e/d8077a96701c53e62c7818c1bc011f6ac490050acbc3db11c07b8fb0a272/interp_engine-1.6.0.tar.gz", hash = "sha256:08952ed57cdc8756c61cee777cd6a1ad51c86db75e561fae42d01f76ab7f97d5", size = 874118, upload-time = "2026-09-04T18:37:33.168Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/cf/21/1b9ecf9485042acf50914323bf3e19a94bdac2496fc2ae21b8f3970a8a5b/interp_engine-1.3.3-py3-none-any.whl", hash = "sha256:8f61b28844ec559a1eae8d3bb41951de4c27e002484163515be7bdb3976830f9", size = 377787, upload-time = "2026-08-21T00:02:37.609Z" },
+    { url = "https://files.pythonhosted.org/packages/3f/ce/59a0615d7cbb2f35060f234372326a6d2b8d26a0065762b2d173f12da946/interp_engine-1.6.0-py3-none-any.whl", hash = "sha256:75e16251b04d44f27e1165edad4b00e32d1e0a838f55476febca2134cd8f7237", size = 439787, upload-time = "2026-09-04T18:37:31.441Z" },
 ]
 
 [[package]]
@@ -1183,7 +1183,7 @@ requires-dist = [
     { name = "circuit-tracer", extras = ["interp-engine"], marker = "extra == 'circuit-tracer'", git = "https://github.com/decoderesearch/circuit-tracer.git?rev=3f0965b6701f96ca498128c7226cb3e13d3c0d5e" },
     { name = "fastapi", specifier = ">=0.115.12" },
     { name = "hf-transfer", specifier = ">=0.1.9,<0.2" },
-    { name = "interp-engine", specifier = "==1.3.3" },
+    { name = "interp-engine", specifier = "==1.6.0" },
     { name = "llamascopium", marker = "extra == 'crm'", git = "https://github.com/OpenMOSS/Language-Model-SAEs.git?rev=cfd202b18b22007f3a293670b7bbac972bd62093" },
     { name = "psutil", specifier = ">=7.0,<8" },
     { name = "pydantic", specifier = ">=2.11.4" },
diff --git a/apps/inference/pyproject.toml b/apps/inference/pyproject.toml
index 14dd77f10..3dedf4e30 100644
--- a/apps/inference/pyproject.toml
+++ b/apps/inference/pyproject.toml
@@ -46,7 +46,10 @@ dependencies = [
   # 1.3.0 renamed the vllm-freeze backend to vllm-static and split generation-only out as
   # vllm-generate; freeze_points/frozen_points became static_points. See AGENT_INTEGRATION.md
   # in the engine repo. This service's call sites moved with it in the same commit.
-  "interp-engine[vllm,quant]==1.3.3",
+  # 1.6.0 is additive on the surface this service calls: speed and memory work (half the
+  # static-buffer VRAM on vllm-static), per-model point fixes for Gemma 4, and a new
+  # `interp_engine.memory` sizer nothing here imports yet.
+  "interp-engine[vllm,quant]==1.6.0",
 ]
 
 [dependency-groups]
diff --git a/apps/inference/uv.lock b/apps/inference/uv.lock
index 73890777a..84ddfaf79 100644
--- a/apps/inference/uv.lock
+++ b/apps/inference/uv.lock
@@ -18,7 +18,8 @@ name = "accelerate"
 version = "1.14.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "numpy", version = "1.26.4", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "packaging" },
@@ -692,7 +693,8 @@ dependencies = [
     { name = "filelock" },
     { name = "fsspec", extra = ["http"] },
     { name = "httpx" },
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "multiprocess" },
     { name = "numpy", version = "1.26.4", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
@@ -1204,26 +1206,43 @@ wheels = [
 name = "hf-xet"
 version = "1.5.1"
 source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version == '3.12.*' and sys_platform != 'linux'",
+    "python_full_version < '3.12' and sys_platform != 'linux'",
+    "python_full_version >= '3.13' and sys_platform != 'linux'",
+]
 sdist = { url = "https://files.pythonhosted.org/packages/4b/2d/57fd21d84d93efb4bd0b962383790e19dd1bc053501b4264c97903b4e83e/hf_xet-1.5.1.tar.gz", hash = "sha256:51ef4500dab3764b41135ee1381a4b62ce56fc54d4c92b719b59e597d6df5bf6", size = 876636, upload-time = "2026-06-08T23:02:53.897Z" }
 wheels = [
     { url = "https://files.pythonhosted.org/packages/64/ee/dd9ba7beae1005e54131b7d45263cc74c8a066d47d354e6d58ae9445a388/hf_xet-1.5.1-cp313-cp313t-macosx_10_12_x86_64.whl", hash = "sha256:dbf48c0d02cf0b2e568944330c60d9120c272dabe013bd892d48e25bc6797577", size = 4069485, upload-time = "2026-06-08T23:02:13.193Z" },
     { url = "https://files.pythonhosted.org/packages/b6/bc/9cae6cfeb4e03070874e73e5c97c66eb90369d3206b6a2b1ef5f96520888/hf_xet-1.5.1-cp313-cp313t-macosx_11_0_arm64.whl", hash = "sha256:e78e4e5192ad2b674c2e1160b651cb9134db974f8ae1835bdfbfb0166b894a43", size = 3838493, upload-time = "2026-06-08T23:02:15.282Z" },
-    { url = "https://files.pythonhosted.org/packages/ba/b4/d5c01e0eb6d9f2ca2dacd84d0d1b71e6cfbb2ef3208c968528e010e9b3d7/hf_xet-1.5.1-cp313-cp313t-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:6f7a04a8ad962422e225bc49fbbac99dc1806764b1f3e54dbd154bffa7593947", size = 4505658, upload-time = "2026-06-08T23:02:17.196Z" },
-    { url = "https://files.pythonhosted.org/packages/76/c5/29a7598c0c6383c523dc22186d577f4e04267a626cd95ae60f67c00bfe66/hf_xet-1.5.1-cp313-cp313t-manylinux_2_28_aarch64.whl", hash = "sha256:d48199c2bf4f8df0adc55d31d1368b6ec0e4d4f45bc86b08038089c23db0bed8", size = 4292822, upload-time = "2026-06-08T23:02:18.608Z" },
-    { url = "https://files.pythonhosted.org/packages/04/9a/dceaf6ca69390126b86ea825fb354b93d01163199070b7bd849225de9468/hf_xet-1.5.1-cp313-cp313t-musllinux_1_2_aarch64.whl", hash = "sha256:97f212a88d14bbf573619a74b7fecb238de77d08fc702e54dec6f78276ca3283", size = 4491255, upload-time = "2026-06-08T23:02:20.124Z" },
-    { url = "https://files.pythonhosted.org/packages/48/a7/e5a7afaacf6c1791fdbeeac42951fb81c3d2bc482992b115dedcc86d963e/hf_xet-1.5.1-cp313-cp313t-musllinux_1_2_x86_64.whl", hash = "sha256:f61e3665892a6c8c5e765395838b8ddf36185da835253d4bc4509a81e49fb342", size = 4711062, upload-time = "2026-06-08T23:02:21.863Z" },
     { url = "https://files.pythonhosted.org/packages/53/49/2802f8433c9742ce281bddc1e65c02c32268ca3098d66828b05e12e45ee2/hf_xet-1.5.1-cp313-cp313t-win_amd64.whl", hash = "sha256:f4ad3ebd4c32dd2b27099d69dc7b2df821e30767e46fb6ee6a0713778243b8ff", size = 4017205, upload-time = "2026-06-08T23:02:23.495Z" },
     { url = "https://files.pythonhosted.org/packages/9e/5a/50c71195b9fb883659f596e7252faf4c18c58e753a9013bdbf9bac5d2250/hf_xet-1.5.1-cp313-cp313t-win_arm64.whl", hash = "sha256:8298485c1e36e7e67cbd01eeb1376619b7af43d4f1ec245caae306f890a8a32d", size = 3845426, upload-time = "2026-06-08T23:02:25.124Z" },
     { url = "https://files.pythonhosted.org/packages/7a/d8/5e54cf37434759d1f4f2ba9b66077ff9d4c4e1f37b6bd7975da5c40d94ab/hf_xet-1.5.1-cp37-abi3-macosx_10_12_x86_64.whl", hash = "sha256:6abd35c3221eff63836618ddfb954dcf84798603f71d8e33e3ed7b04acfdbe6e", size = 4077794, upload-time = "2026-06-08T23:02:40.656Z" },
     { url = "https://files.pythonhosted.org/packages/35/94/4b2ecfbad8f8b04701a23aefb62f540b9137d058b7e1dbef16a32676f0e9/hf_xet-1.5.1-cp37-abi3-macosx_11_0_arm64.whl", hash = "sha256:94e761bbd266bf4c03cee73753916062665ce8365aa40ed321f45afcb934b41e", size = 3845354, upload-time = "2026-06-08T23:02:42.702Z" },
-    { url = "https://files.pythonhosted.org/packages/de/cc/f99f4bc7295023d7bd9ebbfd51f75cc530ca262c1227666268b8208f4b77/hf_xet-1.5.1-cp37-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:892e3a3a3aecc12aded8b93cf4f9cd059282c7de0732f7d55026f3abdf474350", size = 4514864, upload-time = "2026-06-08T23:02:44.497Z" },
-    { url = "https://files.pythonhosted.org/packages/cd/6e/21f7e5a2381278bd3b7b7a5a4d90038518bb6308a0c1daf5d9f8268bb178/hf_xet-1.5.1-cp37-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:a93df2039190502835b1db8cd7e178b0b7b889fe9ab51299d5ced26e0dd879a4", size = 4303784, upload-time = "2026-06-08T23:02:46.203Z" },
-    { url = "https://files.pythonhosted.org/packages/35/0e/f992bb6927ac1cb30ef74e62268f551f338bc32b2191f7c96a44c6f7283e/hf_xet-1.5.1-cp37-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:0c97106032ef70467b4f6bc2d0ccc266d7613ee076afc56516c502f87ce1c4a6", size = 4500703, upload-time = "2026-06-08T23:02:47.628Z" },
-    { url = "https://files.pythonhosted.org/packages/fb/d1/90a498d05447980b977b1669246eeeeae4cfb0ea3e7a286eaba627f91bf9/hf_xet-1.5.1-cp37-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:6208adb15d192b90e4c2ad2a27ed864359b2cb0f2494eb6d7c7f3699ac02e2bf", size = 4719498, upload-time = "2026-06-08T23:02:49.268Z" },
     { url = "https://files.pythonhosted.org/packages/6d/b6/20f99cfe97cc663a711f7b33cc21d4793e51968e9a26125b4afcd77315ba/hf_xet-1.5.1-cp37-abi3-win_amd64.whl", hash = "sha256:f7b3002f95d1c13e24bcb4537baa8f0eb3838957067c91bb4959bc004a6435f5", size = 4026419, upload-time = "2026-06-08T23:02:50.829Z" },
     { url = "https://files.pythonhosted.org/packages/f9/fa/77453694888f03e5a8c8852d1514a0894d8e81c622d39edbaf308ea0dcf4/hf_xet-1.5.1-cp37-abi3-win_arm64.whl", hash = "sha256:93d090b57b211133f6c0dab0205ef5cb6d89162979ba75a74845045cc3063b8e", size = 3855178, upload-time = "2026-06-08T23:02:52.452Z" },
 ]
 
+[[package]]
+name = "hf-xet"
+version = "1.6.0"
+source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+    "python_full_version < '3.12' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version < '3.12' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+    "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+]
+sdist = { url = "https://files.pythonhosted.org/packages/1b/ab/522a2ab67f27971a9d48ca666d4fca85ef7d5282d142e31fd087e27b1bbe/hf_xet-1.6.0.tar.gz", hash = "sha256:2e58454a340b3556dfa4972d5451aff4fba8dd42a236600ba1a1d2b1514f0fef", size = 920527, upload-time = "2026-08-03T22:33:13.243Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/67/4e/a28359bf1c1ecf11eba22123168c138698f7cb576ac678f5a2e16cd5da08/hf_xet-1.6.0-cp38-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:d62671bb130879cef0ee4c9ebe47a14af6c66ec53e6d84dc15936e5ffdfac82f", size = 4464663, upload-time = "2026-08-03T22:33:03.802Z" },
+    { url = "https://files.pythonhosted.org/packages/9a/69/1f0cbc2fb22ae6082d094f743d1b8945a3f36f6089cb95f42b7ee348cda7/hf_xet-1.6.0-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:0e6e21fa3cdfcdcd76748564bf593870a5e013f47d97cf10aed63aa222cff5b7", size = 4262538, upload-time = "2026-08-03T22:33:05.287Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/3a/4f4f2301ade26e404462d3336fa11f7958d914cabbabdd6e03c3c5d5658c/hf_xet-1.6.0-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:4fc74352a17015bd0ee90038bc9efe38db894cde45f268b6712b04fce8cd0acb", size = 4460520, upload-time = "2026-08-03T22:33:06.81Z" },
+    { url = "https://files.pythonhosted.org/packages/ab/5f/311725e2a905534dfee2dcb5b08414f249147f1f12252bfc2bd24caa075c/hf_xet-1.6.0-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:8fb4f71cba6129110c3374a33f919001ff130488fc23553698e34cc1c2a1198c", size = 4675937, upload-time = "2026-08-03T22:33:08.616Z" },
+]
+
 [[package]]
 name = "httpcore"
 version = "1.0.9"
@@ -1294,41 +1313,74 @@ wheels = [
 name = "huggingface-hub"
 version = "1.24.0"
 source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version == '3.12.*' and sys_platform != 'linux'",
+    "python_full_version < '3.12' and sys_platform != 'linux'",
+    "python_full_version >= '3.13' and sys_platform != 'linux'",
+]
 dependencies = [
-    { name = "click" },
-    { name = "filelock" },
-    { name = "fsspec" },
-    { name = "hf-xet", marker = "platform_machine == 'AMD64' or platform_machine == 'aarch64' or platform_machine == 'amd64' or platform_machine == 'arm64' or platform_machine == 'x86_64'" },
-    { name = "httpx" },
-    { name = "packaging" },
-    { name = "pyyaml" },
-    { name = "tqdm" },
-    { name = "typing-extensions" },
+    { name = "click", marker = "sys_platform != 'linux'" },
+    { name = "filelock", marker = "sys_platform != 'linux'" },
+    { name = "fsspec", marker = "sys_platform != 'linux'" },
+    { name = "hf-xet", version = "1.5.1", source = { registry = "https://pypi.org/simple" }, marker = "(platform_machine == 'AMD64' and sys_platform != 'linux') or (platform_machine == 'aarch64' and sys_platform != 'linux') or (platform_machine == 'amd64' and sys_platform != 'linux') or (platform_machine == 'arm64' and sys_platform != 'linux') or (platform_machine == 'x86_64' and sys_platform != 'linux')" },
+    { name = "httpx", marker = "sys_platform != 'linux'" },
+    { name = "packaging", marker = "sys_platform != 'linux'" },
+    { name = "pyyaml", marker = "sys_platform != 'linux'" },
+    { name = "tqdm", marker = "sys_platform != 'linux'" },
+    { name = "typing-extensions", marker = "sys_platform != 'linux'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/df/9b/d3bb4e7d792835daf34dd7091bbc7d7b4e0437d9388f1ea7239cce49f478/huggingface_hub-1.24.0.tar.gz", hash = "sha256:18431ff4daae0749aa9ba102fc952e314c98e1d30ebdec5319d85ca0a83e1ae5", size = 921848, upload-time = "2026-07-17T09:54:01.022Z" }
 wheels = [
     { url = "https://files.pythonhosted.org/packages/5f/c3/aeaaf3911d2529614be18d1c8b5496afc185560e76568063d517283318af/huggingface_hub-1.24.0-py3-none-any.whl", hash = "sha256:6ed4120a84a6beec900640aa7e346bd766a6b7341e41526fef5dc8bd81fb7d59", size = 771904, upload-time = "2026-07-17T09:53:59.106Z" },
 ]
 
+[[package]]
+name = "huggingface-hub"
+version = "1.30.0"
+source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+    "python_full_version < '3.12' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version < '3.12' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+    "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux'",
+    "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux'",
+]
+dependencies = [
+    { name = "click", marker = "sys_platform == 'linux'" },
+    { name = "filelock", marker = "sys_platform == 'linux'" },
+    { name = "fsspec", marker = "sys_platform == 'linux'" },
+    { name = "hf-xet", version = "1.6.0", source = { registry = "https://pypi.org/simple" }, marker = "(platform_machine == 'AMD64' and sys_platform == 'linux') or (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'amd64' and sys_platform == 'linux') or (platform_machine == 'arm64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')" },
+    { name = "httpx", marker = "sys_platform == 'linux'" },
+    { name = "packaging", marker = "sys_platform == 'linux'" },
+    { name = "pyyaml", marker = "sys_platform == 'linux'" },
+    { name = "tqdm", marker = "sys_platform == 'linux'" },
+    { name = "typing-extensions", marker = "sys_platform == 'linux'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/35/97/2eb4abaa5b969ed385066a0496a3823b3ff467fc1082e2202955f1867d60/huggingface_hub-1.30.0.tar.gz", hash = "sha256:e6a6120bc8c8e2723d03648434ee247088cceb55ba7067e7d34d692cad5fdb57", size = 964291, upload-time = "2026-09-03T10:05:14.053Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c3/0e/3e45bbe0dd48f4e56b1d46649d342de853cd1c7e815323472ab62687f153/huggingface_hub-1.30.0-py3-none-any.whl", hash = "sha256:96ae0a8e99a234374a6fe43e989ebd21c04640b91ab2927e7e5773ba1131ca59", size = 796795, upload-time = "2026-09-03T10:05:12.21Z" },
+]
+
 [[package]]
 name = "humming-kernels"
-version = "0.1.10"
+version = "0.1.12"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-bindings", marker = "sys_platform == 'linux'" },
     { name = "jinja2", marker = "sys_platform == 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "nvidia-ml-py", marker = "sys_platform == 'linux'" },
-    { name = "pyelftools", marker = "sys_platform == 'linux'" },
     { name = "safetensors", marker = "sys_platform == 'linux'" },
     { name = "tabulate", marker = "sys_platform == 'linux'" },
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "tqdm", marker = "sys_platform == 'linux'" },
     { name = "triton", marker = "sys_platform == 'linux'" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/51/4f/6977a31451c3f7aa1deaa76506d6cdeb74ef418ad8bcba2e98f7510b26ec/humming_kernels-0.1.10.tar.gz", hash = "sha256:da3e46fb9fc9eba2a9327c2e8135ead68e390c955acd7449f97ee7c71666c8b1", size = 220110, upload-time = "2026-07-02T10:22:57.687Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/93/b9/53797fff4b1b31b4567453a365646d724469b1bb0d76c07f8746617d0cb6/humming_kernels-0.1.12.tar.gz", hash = "sha256:c96c37c95f74379067d2d2d0c5154d12ee9a8cfa29a7c3370a7f3e34f207ca72", size = 259323, upload-time = "2026-07-31T14:28:23.124Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/63/ba/869bc24591d2b4fb0d8da821528072971052a934af077a11f77a0f2b3e79/humming_kernels-0.1.10-py3-none-any.whl", hash = "sha256:4ded0998ff085afeddde70baf93f97c2929969ec3d4a63a52cfec5072bc972b4", size = 184889, upload-time = "2026-07-02T10:22:56.031Z" },
+    { url = "https://files.pythonhosted.org/packages/58/0f/01f871dc26f8d7e1683df9464890d48cfd5c4e46c4a264d7ee0868749197/humming_kernels-0.1.12-py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:988e8b9da41e3679ae2816bdfb51acab53c05cc56254b5ca98e847f4efcf24fd", size = 318019, upload-time = "2026-07-31T14:28:19.925Z" },
+    { url = "https://files.pythonhosted.org/packages/ec/a2/f96912ee7d68f56e61c15555657ffa057b8afbaa21241c63342b5969bb74/humming_kernels-0.1.12-py3-none-manylinux_2_28_x86_64.whl", hash = "sha256:cd3ef712a93f3a9075ea99de2c72bcd3ec89dab3759b3a248d869f5507b60331", size = 322656, upload-time = "2026-07-31T14:28:21.511Z" },
 ]
 
 [package.optional-dependencies]
@@ -1415,7 +1467,7 @@ wheels = [
 
 [[package]]
 name = "interp-engine"
-version = "1.3.3"
+version = "1.6.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "einops" },
@@ -1425,9 +1477,9 @@ dependencies = [
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "transformers" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/10/18/f82b334ef3c81de79eab7423ad273ecd7a1e110f0117b74a96c92b5e9601/interp_engine-1.3.3.tar.gz", hash = "sha256:948fdc71db9119efc484619487d8224c9309e71d05687d3696612d5cc65f3b26", size = 777436, upload-time = "2026-08-21T00:02:39.338Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/42/1e/d8077a96701c53e62c7818c1bc011f6ac490050acbc3db11c07b8fb0a272/interp_engine-1.6.0.tar.gz", hash = "sha256:08952ed57cdc8756c61cee777cd6a1ad51c86db75e561fae42d01f76ab7f97d5", size = 874118, upload-time = "2026-09-04T18:37:33.168Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/cf/21/1b9ecf9485042acf50914323bf3e19a94bdac2496fc2ae21b8f3970a8a5b/interp_engine-1.3.3-py3-none-any.whl", hash = "sha256:8f61b28844ec559a1eae8d3bb41951de4c27e002484163515be7bdb3976830f9", size = 377787, upload-time = "2026-08-21T00:02:37.609Z" },
+    { url = "https://files.pythonhosted.org/packages/3f/ce/59a0615d7cbb2f35060f234372326a6d2b8d26a0065762b2d173f12da946/interp_engine-1.6.0-py3-none-any.whl", hash = "sha256:75e16251b04d44f27e1165edad4b00e32d1e0a838f55476febca2134cd8f7237", size = 439787, upload-time = "2026-09-04T18:37:31.441Z" },
 ]
 
 [package.optional-dependencies]
@@ -1552,7 +1604,8 @@ name = "kernels"
 version = "0.15.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "kernels-data" },
     { name = "packaging" },
     { name = "pyyaml" },
@@ -1976,7 +2029,8 @@ dependencies = [
     { name = "fastapi" },
     { name = "hf-transfer" },
     { name = "huggingface" },
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "interp-engine", extra = ["quant", "vllm"] },
     { name = "numpy", version = "1.26.4", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
@@ -2011,7 +2065,7 @@ requires-dist = [
     { name = "hf-transfer", specifier = ">=0.1.9,<0.2" },
     { name = "huggingface", specifier = ">=0.0.1,<0.0.2" },
     { name = "huggingface-hub", specifier = ">=1.0" },
-    { name = "interp-engine", extras = ["vllm", "quant"], specifier = "==1.3.3" },
+    { name = "interp-engine", extras = ["vllm", "quant"], specifier = "==1.6.0" },
     { name = "numpy", specifier = ">=1.24" },
     { name = "pandas", specifier = ">=2.2.2,<3" },
     { name = "psutil", specifier = ">=5.9.8,<6" },
@@ -2329,14 +2383,14 @@ wheels = [
 
 [[package]]
 name = "nvidia-cutlass-dsl"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "nvidia-cutlass-dsl-libs-base", marker = "sys_platform == 'linux'" },
     { name = "nvidia-cutlass-dsl-libs-cu12", marker = "sys_platform == 'linux'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/8b/1c/fbddb760a0228df87a9e9d1e60b76ecbe6e18035f5853efe0b4563651b2b/nvidia_cutlass_dsl-4.6.0-py3-none-any.whl", hash = "sha256:e3e0e4d8df20d82c8401fa013f4d82021f41daa5fca3d24b55d4a677f2308ca8", size = 10459, upload-time = "2026-07-02T03:23:18.43Z" },
+    { url = "https://files.pythonhosted.org/packages/b2/ec/14e6ecbfed31ec35bbd1bb6965ae61f879370906a8f9c2e851e292704ba4/nvidia_cutlass_dsl-4.6.2-py3-none-any.whl", hash = "sha256:06ac62deb182a852dfb053032cf945bf02b9aa1bf502a18b129d280d7babb26c", size = 10460, upload-time = "2026-08-05T14:39:41.657Z" },
 ]
 
 [package.optional-dependencies]
@@ -2346,7 +2400,7 @@ cu13 = [
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-base"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform == 'linux'" },
@@ -2357,17 +2411,17 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform == 'linux'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/c7/6e/480e2b4c8cfad7271333b44e20e1bcc821632b32b0d5c9c37022ba2e66ee/nvidia_cutlass_dsl_libs_base-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:669200100131a0b8f2876c535ea532c967335936678593c13aced8273ca7523e", size = 3321233, upload-time = "2026-07-02T03:24:44.902Z" },
-    { url = "https://files.pythonhosted.org/packages/3c/88/1f259ffe78178e30a90fbddcec49a567aa077929aec2558f80fcec8dd019/nvidia_cutlass_dsl_libs_base-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:90a3e7a61d110a8ed005aae83869c6e5dca0723e36298297c0780e21db59c016", size = 2826897, upload-time = "2026-07-02T03:25:06.846Z" },
-    { url = "https://files.pythonhosted.org/packages/2c/f8/22653971fcab2a7ed581934f7a2708c9873fa6a8e8eb285422c8eed4ae01/nvidia_cutlass_dsl_libs_base-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:6412572899b1c6d182e516b20f2b0a21874ec88d25234e6040fb2a4381de7a1a", size = 3321728, upload-time = "2026-07-02T03:25:28.888Z" },
-    { url = "https://files.pythonhosted.org/packages/ce/38/e91f66739d2f8711d1a2457e68cd86d6fbae307ce66ce270a405d4dc6dc7/nvidia_cutlass_dsl_libs_base-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:e41cd5db4de4b535c30ae9ca4412b957800a62560019ae91fa51cf3ea89bf254", size = 2824817, upload-time = "2026-07-02T03:25:53.814Z" },
-    { url = "https://files.pythonhosted.org/packages/e9/5c/0a82b9b2fee054788944d0e7a97b5e63ed0d304969c3b5c4168bed86b71c/nvidia_cutlass_dsl_libs_base-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:7b5f5502cc827039f42789e1e2ac9aef7010f4d26ef4b9d66f4ab082da0bcd2c", size = 3321628, upload-time = "2026-07-02T03:26:10.487Z" },
-    { url = "https://files.pythonhosted.org/packages/7f/67/6c21b2d140bbd1ad94a2e22a0e2881457f9e363bdff1b35898a2d7d25aa2/nvidia_cutlass_dsl_libs_base-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:7f27357b87c5c797344cca073f1dcf00232aef427daf161adb3cd87b043e37c9", size = 2824911, upload-time = "2026-07-02T03:26:25.033Z" },
+    { url = "https://files.pythonhosted.org/packages/70/0e/5020da76c7dd3ed74acc6f435e53ca2df967762b2d8408f3e81f146e018a/nvidia_cutlass_dsl_libs_base-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:aae76fcc94c5483279f06848ee470b87417b9b1a4ad8f822a12e2654cbaf513a", size = 3321401, upload-time = "2026-08-05T14:11:32.779Z" },
+    { url = "https://files.pythonhosted.org/packages/12/ce/dfefb22eb438de1e0e6c6627f188449ad3f65448522ba80a12a67bb14715/nvidia_cutlass_dsl_libs_base-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:551097a99537ffdd239678088865ab8d1e273633f0961493509284d8445dd420", size = 2827063, upload-time = "2026-08-05T14:11:54.706Z" },
+    { url = "https://files.pythonhosted.org/packages/aa/f3/4be98f2faa283471e355a42ecbc20956fb2c3b6dbc71861f483be277e99a/nvidia_cutlass_dsl_libs_base-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:e129db7155d56ba4478be098c0f819e37690a6ce5d43a9a9a83bf1300d32c57f", size = 3321894, upload-time = "2026-08-05T14:12:30.139Z" },
+    { url = "https://files.pythonhosted.org/packages/54/1b/cf4a0837054bb4b9dc36cdd86f9d2e7fead3629e3373aaa7f7ec1e5e1542/nvidia_cutlass_dsl_libs_base-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:24dfaddad6077fd0de14eecaf66a72762f2f5f30ae1872231f5bf8b243ac2eac", size = 2824987, upload-time = "2026-08-05T14:12:49.618Z" },
+    { url = "https://files.pythonhosted.org/packages/41/e9/a402e079e926f1fbcf82e30dfcd87aa924aaf91b02db08d57e83b158dbb2/nvidia_cutlass_dsl_libs_base-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:334a9be3c5054286666b14f81c9278075c2007b947ba75f7ee44b94aadf26415", size = 3321794, upload-time = "2026-08-05T14:13:12.501Z" },
+    { url = "https://files.pythonhosted.org/packages/d8/4c/0a6c6e785e8fb2bfc6a0941559d30fbaf6264ae77ee1a26a42e37fd6efdc/nvidia_cutlass_dsl_libs_base-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:896f8a87d0815e59066b0607930c9905a5cf8f5c7a01a423b0093bc876bac4b4", size = 2825075, upload-time = "2026-08-05T14:13:34.746Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-core"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform == 'linux'" },
@@ -2377,12 +2431,12 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform == 'linux'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/84/94/e4e2404ac06a477096ccf8127bf5d391510d36cafb4be86c8c15b4873b0d/nvidia_cutlass_dsl_libs_core-4.6.0-py3-none-any.whl", hash = "sha256:f9ea6d313a03cb11fa177da32e8747ad0cac51358850810f36aa6c4736192c27", size = 767713, upload-time = "2026-07-02T03:23:39.876Z" },
+    { url = "https://files.pythonhosted.org/packages/7c/ef/10c0089234a32f1379e1f213f28f8c7521a9ff0154e3acc58f613d5ef3be/nvidia_cutlass_dsl_libs_core-4.6.2-py3-none-any.whl", hash = "sha256:571d4b46fca1bfbb123d0dc348eaa7fd80af0014ddffc07b849e8195b2364333", size = 772393, upload-time = "2026-08-05T14:09:50.76Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-cu12"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform == 'linux'" },
@@ -2393,17 +2447,17 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform == 'linux'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/d1/c4/ea041a7857b3c7e10e939d787feee59c6c8d2d51a9ae1518684d8894daa1/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:783d5da4aa19a4d11429435419b64264c96d429a1794cbc9df2aa905c1342eee", size = 86992109, upload-time = "2026-07-02T03:28:58.136Z" },
-    { url = "https://files.pythonhosted.org/packages/1e/98/4501c0b4053cacbb4e555d306d891f2426ce7edbb148f6e78376418e0356/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:0479904db0736ab912b2f2db7c6a76cf3c9f6e953f94dc5d667f73c27f18d772", size = 88436391, upload-time = "2026-07-02T03:29:21.26Z" },
-    { url = "https://files.pythonhosted.org/packages/11/38/62def848b65bf067f434df7680c7e8c48519b25bbd3f03f9cdff3606353b/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:87f132ccc30946949868989f3b1b1adaa714ccdf5c636e5379b54909cc29576c", size = 86992102, upload-time = "2026-07-02T03:29:41.47Z" },
-    { url = "https://files.pythonhosted.org/packages/bf/64/f3f8962a9b91dd9368b90e23b2ac81614d6e9df72b55365ec0c216c3f8f9/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:abc341ff0fce40ed0bdadf160f6afac07fb9d01768d4daebd1628c330b3e4210", size = 88436835, upload-time = "2026-07-02T03:30:13.676Z" },
-    { url = "https://files.pythonhosted.org/packages/7c/92/773b79f50ca59ca878a5e6be53d7f407deeef56f0b8000bb8cddc2b66d9e/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:45b72e41d343f6b0c1a98669719e03dca4769d7285ae85b7d2a5292168fc73ec", size = 86992268, upload-time = "2026-07-02T03:30:47.112Z" },
-    { url = "https://files.pythonhosted.org/packages/07/c1/2521ce3d3f46731d0563bf7e5e0fa6b6ea42c31bcb6763cf4bde16d7b3ce/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:22028842dd9c6064a3de7756b301650be77ddebf6f2acd5336bfbcd05aaf4c02", size = 88437265, upload-time = "2026-07-02T03:31:07.225Z" },
+    { url = "https://files.pythonhosted.org/packages/42/26/033408d2f1488e941b3434c21e703ec11c42873166af1eb8348271565782/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:d245dfb3255d95d1ac75063b758acd3e40d5c6699dc98c80f488165260923e11", size = 87011888, upload-time = "2026-08-05T14:18:09.399Z" },
+    { url = "https://files.pythonhosted.org/packages/1a/1b/4bc2842e0b497d9d0fda73976af7640ef742cdb5210e2d143b72c5ec3938/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:4ad5083bf8bcbda83973bdc38a03c2c8452d11fe0afc04b10fc362e23ddd7f8c", size = 88454141, upload-time = "2026-08-05T14:18:31.516Z" },
+    { url = "https://files.pythonhosted.org/packages/ad/b7/019a8f74ad30ad9b6190dd5963d8921fc2b03777316e47260eb822624a0a/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:ea3b96dd6b627fe20637acfa7876ea1adfa8d81a2b106ffff0bb2d042c0c2731", size = 87013242, upload-time = "2026-08-05T14:18:57.703Z" },
+    { url = "https://files.pythonhosted.org/packages/80/6e/70c545a600b52b3f0fbbe01608aefc4e675c924ba886485ad72541f5c88e/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:7f510307369d522da8e7d666557b9b9e7df06b94748bd5dc0198d59dfd0d918a", size = 88454261, upload-time = "2026-08-05T14:21:14.733Z" },
+    { url = "https://files.pythonhosted.org/packages/e1/54/6006a2b4fcd05f32be451fe8881968dcbed526438b4eae9a527534062c79/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:2f3b163060325777a3847954b3a599556b2bff500eca19806a81d1d635bb4149", size = 87012255, upload-time = "2026-08-05T14:21:44.507Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/7a/a08ac0ecdd88bb129a56d95b7c58ea6826e6f5fa00191ef7ecdff8085836/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:84243743382948c83ca627c2dcb106db3844c31d45b8cc2af3d14bc9df0a535e", size = 88453568, upload-time = "2026-08-05T14:22:49.158Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-cu13"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform == 'linux'" },
@@ -2414,12 +2468,12 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform == 'linux'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/15/7c/2b5c8e98511d9f3f778644a31039d3f69763c05a50574a9d31e8a4a3f2ba/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:4876ba55f94f2b1d2285be67ec36e39668cf1ac15b5e7db49830ba5897ea9024", size = 86705152, upload-time = "2026-07-02T03:33:54.958Z" },
-    { url = "https://files.pythonhosted.org/packages/68/b8/a44b389a74f67922e11b888b05db7435228a48a69e3d29b3a5ec579ffbdd/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:7235c5cfd1db3814bf559d6b976beb759dc7298914f4ff2684a91c3071369598", size = 88026175, upload-time = "2026-07-02T03:34:16.054Z" },
-    { url = "https://files.pythonhosted.org/packages/e9/bc/25d542974cc7d2594a22ce1df71c40f8900d44c10aef577cbf5ae7a37e5d/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:8c47899a5778dba4f30de76cce5e22bb5dc2a5bde4979ee1196ec9c6468e80e1", size = 86704408, upload-time = "2026-07-02T03:34:36.929Z" },
-    { url = "https://files.pythonhosted.org/packages/b1/0f/bd8b25e6307764a7bfefa519241d6e417f3ba1c75ce548aa76b712a4fd15/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:4799fabc4bd1f7825ff00101ae0054c5cfb6769aefd6d9694f6da8c0f07e12c3", size = 88026053, upload-time = "2026-07-02T03:34:59.316Z" },
-    { url = "https://files.pythonhosted.org/packages/c9/37/089305ba886ae9ce9359ceba7a8289af744efcaf1b44c85cd3d7c803b9a1/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:f4c0835344efb02d4058f812f452fe33dc18c7f19eb94268860162500cdd2354", size = 86704416, upload-time = "2026-07-02T03:35:22.412Z" },
-    { url = "https://files.pythonhosted.org/packages/59/1d/e39c5b41e5ca59d5a0809fdfb9ac548430ed957a5e806ad85a77d132c15a/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:c3fd95c840748879ece42b7ecc9585205e4228ebde55c04ecbce48d2987d905a", size = 88026053, upload-time = "2026-07-02T03:35:48.41Z" },
+    { url = "https://files.pythonhosted.org/packages/ed/28/5341cc64b3c2c6c697686d64f8e35b3f3fab05055ef69f3380a577f12182/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:3a891048e8227e8c04c09ded2dfbc705bbda5b44a740c409794a67545d86dc1f", size = 86712079, upload-time = "2026-08-05T14:26:43.526Z" },
+    { url = "https://files.pythonhosted.org/packages/14/96/734d780c03b988b0622cd94fec61fecbc83333cd3ed36cede9019fd8fa61/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:e2a3c8f7ba5c98d5e2f6c5231926ceca0cf2c8f7d75dd2e6660394c689b89f7e", size = 88036498, upload-time = "2026-08-05T14:27:05.581Z" },
+    { url = "https://files.pythonhosted.org/packages/46/38/bbe5b8f1ef1a15e7bd1e2ea86686c38f15eeee0a4ba440a5d00d8222cb8e/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:b9543257622ce70d696c6a2bda7e9617ef5b1cc9f29bfe66f4b38d1d1e29a8cb", size = 86712108, upload-time = "2026-08-05T14:27:35.488Z" },
+    { url = "https://files.pythonhosted.org/packages/ca/05/7874c37d50e112177a64a96dd6c7eaeacb1661ebb864e0fe5cfd9f706767/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:a299a623fad75eb752c1a4e447b924e9049ed396b037509c41a48da82d1f340a", size = 88036030, upload-time = "2026-08-05T14:27:58.991Z" },
+    { url = "https://files.pythonhosted.org/packages/33/35/a38b52a8f99d1549c31680deae6a44328f3840e76a1202d5ab0af5b91f58/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:5d215db842c47862232537a6737b0cbf4992318aed24ddb1d4f18c3a43022682", size = 86711155, upload-time = "2026-08-05T14:29:18.118Z" },
+    { url = "https://files.pythonhosted.org/packages/de/b7/75d9470796d4c8fb8e10165825a9cd1e95316bab45af1a6a3dd9555cfe70/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:ad108ba4a6d35c0549fa3860ccbb5a90f005f240d2750ac9f8c3bdaa1f1864cc", size = 88036124, upload-time = "2026-08-05T14:31:27.349Z" },
 ]
 
 [[package]]
@@ -3232,15 +3286,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/77/c1/6e422f34e569cf8e18df68d1939c81c099d2b61e4f7d9621c8a77560799c/pydantic_settings-2.14.2-py3-none-any.whl", hash = "sha256:a20c97b37910b6550d5ea50fbcc2d4187defe58cd57070b73863d069419c9440", size = 61715, upload-time = "2026-06-19T13:44:55.02Z" },
 ]
 
-[[package]]
-name = "pyelftools"
-version = "0.33"
-source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/a3/11/767522582afab1b884d277de0e6e011640cb9d7292a38694b4b1a1df1ae8/pyelftools-0.33.tar.gz", hash = "sha256:660d82dcbeb8e83d1702bd97f223f761625da06111c0cc988eac6b8ab0c1b61f", size = 15068655, upload-time = "2026-05-29T12:56:22.553Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/46/2a/f9697576603dae937727827505a6126a066affb227034e77e6f9068910da/pyelftools-0.33-py3-none-any.whl", hash = "sha256:f215ad5f47d3f1373a21496a6c9e0707c622840d0622f23ff7ce08678b020036", size = 201178, upload-time = "2026-05-29T12:56:20.587Z" },
-]
-
 [[package]]
 name = "pygments"
 version = "2.20.0"
@@ -3441,7 +3486,7 @@ wheels = [
 
 [[package]]
 name = "quack-kernels"
-version = "0.6.1"
+version = "0.6.4"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "apache-tvm-ffi", marker = "sys_platform == 'linux'" },
@@ -3450,9 +3495,9 @@ dependencies = [
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "torch-c-dlpack-ext", marker = "sys_platform == 'linux'" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/01/17/890875f88f4d7da28faec9e6cf0a0cc565715b01474e50501fafc5bc71b4/quack_kernels-0.6.1.tar.gz", hash = "sha256:a694f89c91d137478de523c0227365a331ac9cb66790cfb08baa3dbfaafc71e7", size = 387353, upload-time = "2026-07-05T11:50:19.807Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/75/5c/67c4a0d54cbb8d4d05af73132d620d1900a193b33e0d5ea6a25829b941d2/quack_kernels-0.6.4.tar.gz", hash = "sha256:8bf08a0e1a85aecc892ce84f4b6ed7abb6a44ec7c68b83040abe174bc8c2b936", size = 845485, upload-time = "2026-08-07T23:25:31.94Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/2b/65/a38a30a6ac96a757363a5be9d09cef799640bb143a64ba5a2f4d400d95d9/quack_kernels-0.6.1-py3-none-any.whl", hash = "sha256:266705ea82117e9b1c8a9e44d68a458519f2498d966c0efffd6812120c3995ad", size = 358439, upload-time = "2026-07-05T11:50:18.502Z" },
+    { url = "https://files.pythonhosted.org/packages/28/f5/36ec7e075e90bd92bdf30fee0606f189348c2bbb0af6aebaf7382927c622/quack_kernels-0.6.4-py3-none-any.whl", hash = "sha256:e77c5d1f1299b0b38487fe8737df6c6975daa16bca7f7eb883bd1a74d09e7e78", size = 728458, upload-time = "2026-08-07T23:25:30.549Z" },
 ]
 
 [[package]]
@@ -4133,7 +4178,8 @@ name = "tokenizers"
 version = "0.22.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/73/6f/f80cfef4a312e1fb34baf7d85c72d4411afde10978d4657f8cdd811d3ccc/tokenizers-0.22.2.tar.gz", hash = "sha256:473b83b915e547aa366d1eee11806deaf419e17be16310ac0a14077f1e28f917", size = 372115, upload-time = "2026-01-05T10:45:15.988Z" }
 wheels = [
@@ -4420,7 +4466,7 @@ dependencies = [
     { name = "datasets", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
     { name = "einops", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
     { name = "fancy-einsum", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
-    { name = "huggingface-hub", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
     { name = "jaxtyping", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
     { name = "packaging", marker = "python_full_version < '3.13' and sys_platform == 'linux'" },
@@ -4446,7 +4492,8 @@ name = "transformers"
 version = "5.14.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "numpy", version = "1.26.4", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
     { name = "numpy", version = "2.3.5", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "packaging" },
@@ -4595,7 +4642,7 @@ wheels = [
 
 [[package]]
 name = "vllm"
-version = "0.27.1"
+version = "0.28.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "aiohttp", marker = "sys_platform == 'linux'" },
@@ -4612,6 +4659,7 @@ dependencies = [
     { name = "fastsafetensors", marker = "sys_platform == 'linux'" },
     { name = "filelock", marker = "sys_platform == 'linux'" },
     { name = "flashinfer-python", marker = "sys_platform == 'linux'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" },
     { name = "humming-kernels", extra = ["cu13"], marker = "sys_platform == 'linux'" },
     { name = "ijson", marker = "sys_platform == 'linux'" },
     { name = "jsonschema", marker = "sys_platform == 'linux'" },
@@ -4672,10 +4720,10 @@ dependencies = [
     { name = "watchfiles", marker = "sys_platform == 'linux'" },
     { name = "xgrammar", marker = "(platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'arm64' and sys_platform == 'linux') or (platform_machine == 'ppc64le' and sys_platform == 'linux') or (platform_machine == 's390x' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/97/fa/a3cb3a7baf9f1f1027580b0542e42f655b0f577b8a8d39878c1c0a73e6a9/vllm-0.27.1.tar.gz", hash = "sha256:eec2d54d137ac1e59cb4c39226dfee1943eefc8f4788f5821d7300d6acbdb646", size = 39257426, upload-time = "2026-08-11T10:59:48.04Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/22/39/096d88cfe34ee43a8b6e7c957d1ad6338fa5256474d6ff9a32008c6f8cd6/vllm-0.28.0.tar.gz", hash = "sha256:ac96dd0ec5be9c13f2aa4bfb50498c4727110406f6e4980b58a4097e4d18634a", size = 39991441, upload-time = "2026-08-26T10:08:30.56Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/ff/be/6d309884d11ef02966c06d633e156d2070b986df6fccef25041c4b9510e2/vllm-0.27.1-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:74c1a04548c9016ace8f6c03592edf62517152a0e833f46282556ffbe0796254", size = 307180998, upload-time = "2026-08-11T11:00:37.201Z" },
-    { url = "https://files.pythonhosted.org/packages/8f/73/20ad0cc655a5739d95342075c9afab1a25338d4b584fc71548c0b362404b/vllm-0.27.1-cp38-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:98e9fc2a1ed8549a733c9d1b242e2002b82367da9e29e37801761438cb3a2670", size = 312887766, upload-time = "2026-08-11T11:00:11.706Z" },
+    { url = "https://files.pythonhosted.org/packages/b4/31/007ac11dc756f7e69600b90b08e39f573eed9aaee6258198f9d9a2b06355/vllm-0.28.0-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:817b8181f7f61b4a62dc1d5d9ab39f2bfb60a6cb86c29879a78a147b85756787", size = 308873410, upload-time = "2026-08-26T10:08:13.738Z" },
+    { url = "https://files.pythonhosted.org/packages/87/d7/97f6ecc2ae883e601e08d7cef87cb54ececeefcfe6b5e12d5d92f8d06d6b/vllm-0.28.0-cp38-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:addb0ffdaafd8155d75e9b3f5ddb3da28fdee9e8a7097ede91f7db2e9e1a3889", size = 314500103, upload-time = "2026-08-26T10:07:25.091Z" },
 ]
 
 [[package]]
diff --git a/apps/nla/pyproject.toml b/apps/nla/pyproject.toml
index 2eda572eb..3663bb8e2 100644
--- a/apps/nla/pyproject.toml
+++ b/apps/nla/pyproject.toml
@@ -23,7 +23,7 @@ dependencies = [
     # the vendored sglang stack. The vLLM version pin lives in the engine.
     # Pinned exactly, not floored: the engine is first-party and moves in lockstep with this
     # service, so a relock for an unrelated dependency must not swap it. Bump deliberately.
-    "interp-engine[vllm]==1.3.3",
+    "interp-engine[vllm]==1.6.0",
     # Weight-only quantization for the source model and reconstructor backbone
     # (--fp8-source / --fp8-reconstructor / --int4-reconstructor). Used to arrive
     # transitively through `sglang[all]` (which pinned torchao==0.9.0); with sglang
diff --git a/apps/nla/uv.lock b/apps/nla/uv.lock
index 74de5e15e..b8ab1dd0f 100644
--- a/apps/nla/uv.lock
+++ b/apps/nla/uv.lock
@@ -15,7 +15,8 @@ name = "accelerate"
 version = "1.13.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'darwin'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13'" },
     { name = "numpy", version = "2.4.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.13'" },
     { name = "packaging" },
@@ -999,25 +1000,55 @@ wheels = [
 name = "huggingface-hub"
 version = "1.24.0"
 source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version >= '3.13' and sys_platform == 'darwin'",
+    "python_full_version == '3.12.*' and sys_platform == 'darwin'",
+    "python_full_version < '3.12' and sys_platform == 'darwin'",
+]
 dependencies = [
-    { name = "click" },
-    { name = "filelock" },
-    { name = "fsspec" },
-    { name = "hf-xet", marker = "platform_machine == 'AMD64' or platform_machine == 'aarch64' or platform_machine == 'amd64' or platform_machine == 'arm64' or platform_machine == 'x86_64'" },
-    { name = "httpx" },
-    { name = "packaging" },
-    { name = "pyyaml" },
-    { name = "tqdm" },
-    { name = "typing-extensions" },
+    { name = "click", marker = "sys_platform == 'darwin'" },
+    { name = "filelock", marker = "sys_platform == 'darwin'" },
+    { name = "fsspec", marker = "sys_platform == 'darwin'" },
+    { name = "hf-xet", marker = "(platform_machine == 'AMD64' and sys_platform == 'darwin') or (platform_machine == 'aarch64' and sys_platform == 'darwin') or (platform_machine == 'amd64' and sys_platform == 'darwin') or (platform_machine == 'arm64' and sys_platform == 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
+    { name = "httpx", marker = "sys_platform == 'darwin'" },
+    { name = "packaging", marker = "sys_platform == 'darwin'" },
+    { name = "pyyaml", marker = "sys_platform == 'darwin'" },
+    { name = "tqdm", marker = "sys_platform == 'darwin'" },
+    { name = "typing-extensions", marker = "sys_platform == 'darwin'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/df/9b/d3bb4e7d792835daf34dd7091bbc7d7b4e0437d9388f1ea7239cce49f478/huggingface_hub-1.24.0.tar.gz", hash = "sha256:18431ff4daae0749aa9ba102fc952e314c98e1d30ebdec5319d85ca0a83e1ae5", size = 921848, upload-time = "2026-07-17T09:54:01.022Z" }
 wheels = [
     { url = "https://files.pythonhosted.org/packages/5f/c3/aeaaf3911d2529614be18d1c8b5496afc185560e76568063d517283318af/huggingface_hub-1.24.0-py3-none-any.whl", hash = "sha256:6ed4120a84a6beec900640aa7e346bd766a6b7341e41526fef5dc8bd81fb7d59", size = 771904, upload-time = "2026-07-17T09:53:59.106Z" },
 ]
 
+[[package]]
+name = "huggingface-hub"
+version = "1.30.0"
+source = { registry = "https://pypi.org/simple" }
+resolution-markers = [
+    "python_full_version >= '3.13' and sys_platform != 'darwin'",
+    "python_full_version == '3.12.*' and sys_platform != 'darwin'",
+    "python_full_version < '3.12' and sys_platform != 'darwin'",
+]
+dependencies = [
+    { name = "click", marker = "sys_platform != 'darwin'" },
+    { name = "filelock", marker = "sys_platform != 'darwin'" },
+    { name = "fsspec", marker = "sys_platform != 'darwin'" },
+    { name = "hf-xet", marker = "(platform_machine == 'AMD64' and sys_platform != 'darwin') or (platform_machine == 'aarch64' and sys_platform != 'darwin') or (platform_machine == 'amd64' and sys_platform != 'darwin') or (platform_machine == 'arm64' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform != 'darwin')" },
+    { name = "httpx", marker = "sys_platform != 'darwin'" },
+    { name = "packaging", marker = "sys_platform != 'darwin'" },
+    { name = "pyyaml", marker = "sys_platform != 'darwin'" },
+    { name = "tqdm", marker = "sys_platform != 'darwin'" },
+    { name = "typing-extensions", marker = "sys_platform != 'darwin'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/35/97/2eb4abaa5b969ed385066a0496a3823b3ff467fc1082e2202955f1867d60/huggingface_hub-1.30.0.tar.gz", hash = "sha256:e6a6120bc8c8e2723d03648434ee247088cceb55ba7067e7d34d692cad5fdb57", size = 964291, upload-time = "2026-09-03T10:05:14.053Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c3/0e/3e45bbe0dd48f4e56b1d46649d342de853cd1c7e815323472ab62687f153/huggingface_hub-1.30.0-py3-none-any.whl", hash = "sha256:96ae0a8e99a234374a6fe43e989ebd21c04640b91ab2927e7e5773ba1131ca59", size = 796795, upload-time = "2026-09-03T10:05:12.21Z" },
+]
+
 [[package]]
 name = "humming-kernels"
-version = "0.1.10"
+version = "0.1.12"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-bindings", marker = "sys_platform != 'darwin'" },
@@ -1025,16 +1056,16 @@ dependencies = [
     { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13' and sys_platform != 'darwin'" },
     { name = "numpy", version = "2.4.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.13' and sys_platform != 'darwin'" },
     { name = "nvidia-ml-py", marker = "sys_platform != 'darwin'" },
-    { name = "pyelftools", marker = "sys_platform != 'darwin'" },
     { name = "safetensors", marker = "sys_platform != 'darwin'" },
     { name = "tabulate", marker = "sys_platform != 'darwin'" },
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "tqdm", marker = "sys_platform != 'darwin'" },
     { name = "triton", marker = "sys_platform != 'darwin'" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/51/4f/6977a31451c3f7aa1deaa76506d6cdeb74ef418ad8bcba2e98f7510b26ec/humming_kernels-0.1.10.tar.gz", hash = "sha256:da3e46fb9fc9eba2a9327c2e8135ead68e390c955acd7449f97ee7c71666c8b1", size = 220110, upload-time = "2026-07-02T10:22:57.687Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/93/b9/53797fff4b1b31b4567453a365646d724469b1bb0d76c07f8746617d0cb6/humming_kernels-0.1.12.tar.gz", hash = "sha256:c96c37c95f74379067d2d2d0c5154d12ee9a8cfa29a7c3370a7f3e34f207ca72", size = 259323, upload-time = "2026-07-31T14:28:23.124Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/63/ba/869bc24591d2b4fb0d8da821528072971052a934af077a11f77a0f2b3e79/humming_kernels-0.1.10-py3-none-any.whl", hash = "sha256:4ded0998ff085afeddde70baf93f97c2929969ec3d4a63a52cfec5072bc972b4", size = 184889, upload-time = "2026-07-02T10:22:56.031Z" },
+    { url = "https://files.pythonhosted.org/packages/58/0f/01f871dc26f8d7e1683df9464890d48cfd5c4e46c4a264d7ee0868749197/humming_kernels-0.1.12-py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:988e8b9da41e3679ae2816bdfb51acab53c05cc56254b5ca98e847f4efcf24fd", size = 318019, upload-time = "2026-07-31T14:28:19.925Z" },
+    { url = "https://files.pythonhosted.org/packages/ec/a2/f96912ee7d68f56e61c15555657ffa057b8afbaa21241c63342b5969bb74/humming_kernels-0.1.12-py3-none-manylinux_2_28_x86_64.whl", hash = "sha256:cd3ef712a93f3a9075ea99de2c72bcd3ec89dab3759b3a248d869f5507b60331", size = 322656, upload-time = "2026-07-31T14:28:21.511Z" },
 ]
 
 [package.optional-dependencies]
@@ -1103,7 +1134,7 @@ wheels = [
 
 [[package]]
 name = "interp-engine"
-version = "1.3.3"
+version = "1.6.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "einops" },
@@ -1113,9 +1144,9 @@ dependencies = [
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "transformers" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/10/18/f82b334ef3c81de79eab7423ad273ecd7a1e110f0117b74a96c92b5e9601/interp_engine-1.3.3.tar.gz", hash = "sha256:948fdc71db9119efc484619487d8224c9309e71d05687d3696612d5cc65f3b26", size = 777436, upload-time = "2026-08-21T00:02:39.338Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/42/1e/d8077a96701c53e62c7818c1bc011f6ac490050acbc3db11c07b8fb0a272/interp_engine-1.6.0.tar.gz", hash = "sha256:08952ed57cdc8756c61cee777cd6a1ad51c86db75e561fae42d01f76ab7f97d5", size = 874118, upload-time = "2026-09-04T18:37:33.168Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/cf/21/1b9ecf9485042acf50914323bf3e19a94bdac2496fc2ae21b8f3970a8a5b/interp_engine-1.3.3-py3-none-any.whl", hash = "sha256:8f61b28844ec559a1eae8d3bb41951de4c27e002484163515be7bdb3976830f9", size = 377787, upload-time = "2026-08-21T00:02:37.609Z" },
+    { url = "https://files.pythonhosted.org/packages/3f/ce/59a0615d7cbb2f35060f234372326a6d2b8d26a0065762b2d173f12da946/interp_engine-1.6.0-py3-none-any.whl", hash = "sha256:75e16251b04d44f27e1165edad4b00e32d1e0a838f55476febca2134cd8f7237", size = 439787, upload-time = "2026-09-04T18:37:31.441Z" },
 ]
 
 [package.optional-dependencies]
@@ -1573,7 +1604,8 @@ source = { virtual = "." }
 dependencies = [
     { name = "accelerate" },
     { name = "fastapi" },
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'darwin'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "interp-engine", extra = ["vllm"] },
     { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13'" },
     { name = "numpy", version = "2.4.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.13'" },
@@ -1602,7 +1634,7 @@ requires-dist = [
     { name = "accelerate", specifier = ">=1.13.0" },
     { name = "fastapi" },
     { name = "huggingface-hub" },
-    { name = "interp-engine", extras = ["vllm"], specifier = "==1.3.3" },
+    { name = "interp-engine", extras = ["vllm"], specifier = "==1.6.0" },
     { name = "numpy" },
     { name = "orjson" },
     { name = "python-dotenv" },
@@ -1939,14 +1971,14 @@ wheels = [
 
 [[package]]
 name = "nvidia-cutlass-dsl"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "nvidia-cutlass-dsl-libs-base", marker = "sys_platform != 'darwin'" },
     { name = "nvidia-cutlass-dsl-libs-cu12", marker = "sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/8b/1c/fbddb760a0228df87a9e9d1e60b76ecbe6e18035f5853efe0b4563651b2b/nvidia_cutlass_dsl-4.6.0-py3-none-any.whl", hash = "sha256:e3e0e4d8df20d82c8401fa013f4d82021f41daa5fca3d24b55d4a677f2308ca8", size = 10459, upload-time = "2026-07-02T03:23:18.43Z" },
+    { url = "https://files.pythonhosted.org/packages/b2/ec/14e6ecbfed31ec35bbd1bb6965ae61f879370906a8f9c2e851e292704ba4/nvidia_cutlass_dsl-4.6.2-py3-none-any.whl", hash = "sha256:06ac62deb182a852dfb053032cf945bf02b9aa1bf502a18b129d280d7babb26c", size = 10460, upload-time = "2026-08-05T14:39:41.657Z" },
 ]
 
 [package.optional-dependencies]
@@ -1956,7 +1988,7 @@ cu13 = [
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-base"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform != 'darwin'" },
@@ -1968,17 +2000,17 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/c7/6e/480e2b4c8cfad7271333b44e20e1bcc821632b32b0d5c9c37022ba2e66ee/nvidia_cutlass_dsl_libs_base-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:669200100131a0b8f2876c535ea532c967335936678593c13aced8273ca7523e", size = 3321233, upload-time = "2026-07-02T03:24:44.902Z" },
-    { url = "https://files.pythonhosted.org/packages/3c/88/1f259ffe78178e30a90fbddcec49a567aa077929aec2558f80fcec8dd019/nvidia_cutlass_dsl_libs_base-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:90a3e7a61d110a8ed005aae83869c6e5dca0723e36298297c0780e21db59c016", size = 2826897, upload-time = "2026-07-02T03:25:06.846Z" },
-    { url = "https://files.pythonhosted.org/packages/2c/f8/22653971fcab2a7ed581934f7a2708c9873fa6a8e8eb285422c8eed4ae01/nvidia_cutlass_dsl_libs_base-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:6412572899b1c6d182e516b20f2b0a21874ec88d25234e6040fb2a4381de7a1a", size = 3321728, upload-time = "2026-07-02T03:25:28.888Z" },
-    { url = "https://files.pythonhosted.org/packages/ce/38/e91f66739d2f8711d1a2457e68cd86d6fbae307ce66ce270a405d4dc6dc7/nvidia_cutlass_dsl_libs_base-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:e41cd5db4de4b535c30ae9ca4412b957800a62560019ae91fa51cf3ea89bf254", size = 2824817, upload-time = "2026-07-02T03:25:53.814Z" },
-    { url = "https://files.pythonhosted.org/packages/e9/5c/0a82b9b2fee054788944d0e7a97b5e63ed0d304969c3b5c4168bed86b71c/nvidia_cutlass_dsl_libs_base-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:7b5f5502cc827039f42789e1e2ac9aef7010f4d26ef4b9d66f4ab082da0bcd2c", size = 3321628, upload-time = "2026-07-02T03:26:10.487Z" },
-    { url = "https://files.pythonhosted.org/packages/7f/67/6c21b2d140bbd1ad94a2e22a0e2881457f9e363bdff1b35898a2d7d25aa2/nvidia_cutlass_dsl_libs_base-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:7f27357b87c5c797344cca073f1dcf00232aef427daf161adb3cd87b043e37c9", size = 2824911, upload-time = "2026-07-02T03:26:25.033Z" },
+    { url = "https://files.pythonhosted.org/packages/70/0e/5020da76c7dd3ed74acc6f435e53ca2df967762b2d8408f3e81f146e018a/nvidia_cutlass_dsl_libs_base-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:aae76fcc94c5483279f06848ee470b87417b9b1a4ad8f822a12e2654cbaf513a", size = 3321401, upload-time = "2026-08-05T14:11:32.779Z" },
+    { url = "https://files.pythonhosted.org/packages/12/ce/dfefb22eb438de1e0e6c6627f188449ad3f65448522ba80a12a67bb14715/nvidia_cutlass_dsl_libs_base-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:551097a99537ffdd239678088865ab8d1e273633f0961493509284d8445dd420", size = 2827063, upload-time = "2026-08-05T14:11:54.706Z" },
+    { url = "https://files.pythonhosted.org/packages/aa/f3/4be98f2faa283471e355a42ecbc20956fb2c3b6dbc71861f483be277e99a/nvidia_cutlass_dsl_libs_base-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:e129db7155d56ba4478be098c0f819e37690a6ce5d43a9a9a83bf1300d32c57f", size = 3321894, upload-time = "2026-08-05T14:12:30.139Z" },
+    { url = "https://files.pythonhosted.org/packages/54/1b/cf4a0837054bb4b9dc36cdd86f9d2e7fead3629e3373aaa7f7ec1e5e1542/nvidia_cutlass_dsl_libs_base-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:24dfaddad6077fd0de14eecaf66a72762f2f5f30ae1872231f5bf8b243ac2eac", size = 2824987, upload-time = "2026-08-05T14:12:49.618Z" },
+    { url = "https://files.pythonhosted.org/packages/41/e9/a402e079e926f1fbcf82e30dfcd87aa924aaf91b02db08d57e83b158dbb2/nvidia_cutlass_dsl_libs_base-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:334a9be3c5054286666b14f81c9278075c2007b947ba75f7ee44b94aadf26415", size = 3321794, upload-time = "2026-08-05T14:13:12.501Z" },
+    { url = "https://files.pythonhosted.org/packages/d8/4c/0a6c6e785e8fb2bfc6a0941559d30fbaf6264ae77ee1a26a42e37fd6efdc/nvidia_cutlass_dsl_libs_base-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:896f8a87d0815e59066b0607930c9905a5cf8f5c7a01a423b0093bc876bac4b4", size = 2825075, upload-time = "2026-08-05T14:13:34.746Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-core"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform != 'darwin'" },
@@ -1989,12 +2021,12 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/84/94/e4e2404ac06a477096ccf8127bf5d391510d36cafb4be86c8c15b4873b0d/nvidia_cutlass_dsl_libs_core-4.6.0-py3-none-any.whl", hash = "sha256:f9ea6d313a03cb11fa177da32e8747ad0cac51358850810f36aa6c4736192c27", size = 767713, upload-time = "2026-07-02T03:23:39.876Z" },
+    { url = "https://files.pythonhosted.org/packages/7c/ef/10c0089234a32f1379e1f213f28f8c7521a9ff0154e3acc58f613d5ef3be/nvidia_cutlass_dsl_libs_core-4.6.2-py3-none-any.whl", hash = "sha256:571d4b46fca1bfbb123d0dc348eaa7fd80af0014ddffc07b849e8195b2364333", size = 772393, upload-time = "2026-08-05T14:09:50.76Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-cu12"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform != 'darwin'" },
@@ -2006,17 +2038,17 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/d1/c4/ea041a7857b3c7e10e939d787feee59c6c8d2d51a9ae1518684d8894daa1/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:783d5da4aa19a4d11429435419b64264c96d429a1794cbc9df2aa905c1342eee", size = 86992109, upload-time = "2026-07-02T03:28:58.136Z" },
-    { url = "https://files.pythonhosted.org/packages/1e/98/4501c0b4053cacbb4e555d306d891f2426ce7edbb148f6e78376418e0356/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:0479904db0736ab912b2f2db7c6a76cf3c9f6e953f94dc5d667f73c27f18d772", size = 88436391, upload-time = "2026-07-02T03:29:21.26Z" },
-    { url = "https://files.pythonhosted.org/packages/11/38/62def848b65bf067f434df7680c7e8c48519b25bbd3f03f9cdff3606353b/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:87f132ccc30946949868989f3b1b1adaa714ccdf5c636e5379b54909cc29576c", size = 86992102, upload-time = "2026-07-02T03:29:41.47Z" },
-    { url = "https://files.pythonhosted.org/packages/bf/64/f3f8962a9b91dd9368b90e23b2ac81614d6e9df72b55365ec0c216c3f8f9/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:abc341ff0fce40ed0bdadf160f6afac07fb9d01768d4daebd1628c330b3e4210", size = 88436835, upload-time = "2026-07-02T03:30:13.676Z" },
-    { url = "https://files.pythonhosted.org/packages/7c/92/773b79f50ca59ca878a5e6be53d7f407deeef56f0b8000bb8cddc2b66d9e/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:45b72e41d343f6b0c1a98669719e03dca4769d7285ae85b7d2a5292168fc73ec", size = 86992268, upload-time = "2026-07-02T03:30:47.112Z" },
-    { url = "https://files.pythonhosted.org/packages/07/c1/2521ce3d3f46731d0563bf7e5e0fa6b6ea42c31bcb6763cf4bde16d7b3ce/nvidia_cutlass_dsl_libs_cu12-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:22028842dd9c6064a3de7756b301650be77ddebf6f2acd5336bfbcd05aaf4c02", size = 88437265, upload-time = "2026-07-02T03:31:07.225Z" },
+    { url = "https://files.pythonhosted.org/packages/42/26/033408d2f1488e941b3434c21e703ec11c42873166af1eb8348271565782/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:d245dfb3255d95d1ac75063b758acd3e40d5c6699dc98c80f488165260923e11", size = 87011888, upload-time = "2026-08-05T14:18:09.399Z" },
+    { url = "https://files.pythonhosted.org/packages/1a/1b/4bc2842e0b497d9d0fda73976af7640ef742cdb5210e2d143b72c5ec3938/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:4ad5083bf8bcbda83973bdc38a03c2c8452d11fe0afc04b10fc362e23ddd7f8c", size = 88454141, upload-time = "2026-08-05T14:18:31.516Z" },
+    { url = "https://files.pythonhosted.org/packages/ad/b7/019a8f74ad30ad9b6190dd5963d8921fc2b03777316e47260eb822624a0a/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:ea3b96dd6b627fe20637acfa7876ea1adfa8d81a2b106ffff0bb2d042c0c2731", size = 87013242, upload-time = "2026-08-05T14:18:57.703Z" },
+    { url = "https://files.pythonhosted.org/packages/80/6e/70c545a600b52b3f0fbbe01608aefc4e675c924ba886485ad72541f5c88e/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:7f510307369d522da8e7d666557b9b9e7df06b94748bd5dc0198d59dfd0d918a", size = 88454261, upload-time = "2026-08-05T14:21:14.733Z" },
+    { url = "https://files.pythonhosted.org/packages/e1/54/6006a2b4fcd05f32be451fe8881968dcbed526438b4eae9a527534062c79/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:2f3b163060325777a3847954b3a599556b2bff500eca19806a81d1d635bb4149", size = 87012255, upload-time = "2026-08-05T14:21:44.507Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/7a/a08ac0ecdd88bb129a56d95b7c58ea6826e6f5fa00191ef7ecdff8085836/nvidia_cutlass_dsl_libs_cu12-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:84243743382948c83ca627c2dcb106db3844c31d45b8cc2af3d14bc9df0a535e", size = 88453568, upload-time = "2026-08-05T14:22:49.158Z" },
 ]
 
 [[package]]
 name = "nvidia-cutlass-dsl-libs-cu13"
-version = "4.6.0"
+version = "4.6.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cuda-python", marker = "sys_platform != 'darwin'" },
@@ -2028,12 +2060,12 @@ dependencies = [
     { name = "typing-extensions", marker = "sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/15/7c/2b5c8e98511d9f3f778644a31039d3f69763c05a50574a9d31e8a4a3f2ba/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:4876ba55f94f2b1d2285be67ec36e39668cf1ac15b5e7db49830ba5897ea9024", size = 86705152, upload-time = "2026-07-02T03:33:54.958Z" },
-    { url = "https://files.pythonhosted.org/packages/68/b8/a44b389a74f67922e11b888b05db7435228a48a69e3d29b3a5ec579ffbdd/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:7235c5cfd1db3814bf559d6b976beb759dc7298914f4ff2684a91c3071369598", size = 88026175, upload-time = "2026-07-02T03:34:16.054Z" },
-    { url = "https://files.pythonhosted.org/packages/e9/bc/25d542974cc7d2594a22ce1df71c40f8900d44c10aef577cbf5ae7a37e5d/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:8c47899a5778dba4f30de76cce5e22bb5dc2a5bde4979ee1196ec9c6468e80e1", size = 86704408, upload-time = "2026-07-02T03:34:36.929Z" },
-    { url = "https://files.pythonhosted.org/packages/b1/0f/bd8b25e6307764a7bfefa519241d6e417f3ba1c75ce548aa76b712a4fd15/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:4799fabc4bd1f7825ff00101ae0054c5cfb6769aefd6d9694f6da8c0f07e12c3", size = 88026053, upload-time = "2026-07-02T03:34:59.316Z" },
-    { url = "https://files.pythonhosted.org/packages/c9/37/089305ba886ae9ce9359ceba7a8289af744efcaf1b44c85cd3d7c803b9a1/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:f4c0835344efb02d4058f812f452fe33dc18c7f19eb94268860162500cdd2354", size = 86704416, upload-time = "2026-07-02T03:35:22.412Z" },
-    { url = "https://files.pythonhosted.org/packages/59/1d/e39c5b41e5ca59d5a0809fdfb9ac548430ed957a5e806ad85a77d132c15a/nvidia_cutlass_dsl_libs_cu13-4.6.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:c3fd95c840748879ece42b7ecc9585205e4228ebde55c04ecbce48d2987d905a", size = 88026053, upload-time = "2026-07-02T03:35:48.41Z" },
+    { url = "https://files.pythonhosted.org/packages/ed/28/5341cc64b3c2c6c697686d64f8e35b3f3fab05055ef69f3380a577f12182/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:3a891048e8227e8c04c09ded2dfbc705bbda5b44a740c409794a67545d86dc1f", size = 86712079, upload-time = "2026-08-05T14:26:43.526Z" },
+    { url = "https://files.pythonhosted.org/packages/14/96/734d780c03b988b0622cd94fec61fecbc83333cd3ed36cede9019fd8fa61/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:e2a3c8f7ba5c98d5e2f6c5231926ceca0cf2c8f7d75dd2e6660394c689b89f7e", size = 88036498, upload-time = "2026-08-05T14:27:05.581Z" },
+    { url = "https://files.pythonhosted.org/packages/46/38/bbe5b8f1ef1a15e7bd1e2ea86686c38f15eeee0a4ba440a5d00d8222cb8e/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:b9543257622ce70d696c6a2bda7e9617ef5b1cc9f29bfe66f4b38d1d1e29a8cb", size = 86712108, upload-time = "2026-08-05T14:27:35.488Z" },
+    { url = "https://files.pythonhosted.org/packages/ca/05/7874c37d50e112177a64a96dd6c7eaeacb1661ebb864e0fe5cfd9f706767/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:a299a623fad75eb752c1a4e447b924e9049ed396b037509c41a48da82d1f340a", size = 88036030, upload-time = "2026-08-05T14:27:58.991Z" },
+    { url = "https://files.pythonhosted.org/packages/33/35/a38b52a8f99d1549c31680deae6a44328f3840e76a1202d5ab0af5b91f58/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:5d215db842c47862232537a6737b0cbf4992318aed24ddb1d4f18c3a43022682", size = 86711155, upload-time = "2026-08-05T14:29:18.118Z" },
+    { url = "https://files.pythonhosted.org/packages/de/b7/75d9470796d4c8fb8e10165825a9cd1e95316bab45af1a6a3dd9555cfe70/nvidia_cutlass_dsl_libs_cu13-4.6.2-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:ad108ba4a6d35c0549fa3860ccbb5a90f005f240d2750ac9f8c3bdaa1f1864cc", size = 88036124, upload-time = "2026-08-05T14:31:27.349Z" },
 ]
 
 [[package]]
@@ -2742,15 +2774,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/77/c1/6e422f34e569cf8e18df68d1939c81c099d2b61e4f7d9621c8a77560799c/pydantic_settings-2.14.2-py3-none-any.whl", hash = "sha256:a20c97b37910b6550d5ea50fbcc2d4187defe58cd57070b73863d069419c9440", size = 61715, upload-time = "2026-06-19T13:44:55.02Z" },
 ]
 
-[[package]]
-name = "pyelftools"
-version = "0.33"
-source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/a3/11/767522582afab1b884d277de0e6e011640cb9d7292a38694b4b1a1df1ae8/pyelftools-0.33.tar.gz", hash = "sha256:660d82dcbeb8e83d1702bd97f223f761625da06111c0cc988eac6b8ab0c1b61f", size = 15068655, upload-time = "2026-05-29T12:56:22.553Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/46/2a/f9697576603dae937727827505a6126a066affb227034e77e6f9068910da/pyelftools-0.33-py3-none-any.whl", hash = "sha256:f215ad5f47d3f1373a21496a6c9e0707c622840d0622f23ff7ce08678b020036", size = 201178, upload-time = "2026-05-29T12:56:20.587Z" },
-]
-
 [[package]]
 name = "pygments"
 version = "2.19.2"
@@ -2916,7 +2939,7 @@ wheels = [
 
 [[package]]
 name = "quack-kernels"
-version = "0.6.1"
+version = "0.6.4"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "apache-tvm-ffi", marker = "sys_platform != 'darwin'" },
@@ -2925,9 +2948,9 @@ dependencies = [
     { name = "torch", version = "2.13.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "torch-c-dlpack-ext", marker = "sys_platform != 'darwin'" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/01/17/890875f88f4d7da28faec9e6cf0a0cc565715b01474e50501fafc5bc71b4/quack_kernels-0.6.1.tar.gz", hash = "sha256:a694f89c91d137478de523c0227365a331ac9cb66790cfb08baa3dbfaafc71e7", size = 387353, upload-time = "2026-07-05T11:50:19.807Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/75/5c/67c4a0d54cbb8d4d05af73132d620d1900a193b33e0d5ea6a25829b941d2/quack_kernels-0.6.4.tar.gz", hash = "sha256:8bf08a0e1a85aecc892ce84f4b6ed7abb6a44ec7c68b83040abe174bc8c2b936", size = 845485, upload-time = "2026-08-07T23:25:31.94Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/2b/65/a38a30a6ac96a757363a5be9d09cef799640bb143a64ba5a2f4d400d95d9/quack_kernels-0.6.1-py3-none-any.whl", hash = "sha256:266705ea82117e9b1c8a9e44d68a458519f2498d966c0efffd6812120c3995ad", size = 358439, upload-time = "2026-07-05T11:50:18.502Z" },
+    { url = "https://files.pythonhosted.org/packages/28/f5/36ec7e075e90bd92bdf30fee0606f189348c2bbb0af6aebaf7382927c622/quack_kernels-0.6.4-py3-none-any.whl", hash = "sha256:e77c5d1f1299b0b38487fe8737df6c6975daa16bca7f7eb883bd1a74d09e7e78", size = 728458, upload-time = "2026-08-07T23:25:30.549Z" },
 ]
 
 [[package]]
@@ -3432,7 +3455,8 @@ name = "tokenizers"
 version = "0.22.2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'darwin'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/73/6f/f80cfef4a312e1fb34baf7d85c72d4411afde10978d4657f8cdd811d3ccc/tokenizers-0.22.2.tar.gz", hash = "sha256:473b83b915e547aa366d1eee11806deaf419e17be16310ac0a14077f1e28f917", size = 372115, upload-time = "2026-01-05T10:45:15.988Z" }
 wheels = [
@@ -3631,7 +3655,8 @@ name = "transformers"
 version = "5.14.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
+    { name = "huggingface-hub", version = "1.24.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'darwin'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.13'" },
     { name = "numpy", version = "2.4.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.13'" },
     { name = "packaging" },
@@ -3757,7 +3782,7 @@ wheels = [
 
 [[package]]
 name = "vllm"
-version = "0.27.1"
+version = "0.28.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "aiohttp", marker = "sys_platform != 'darwin'" },
@@ -3774,6 +3799,7 @@ dependencies = [
     { name = "fastsafetensors", marker = "sys_platform != 'darwin'" },
     { name = "filelock", marker = "sys_platform != 'darwin'" },
     { name = "flashinfer-python", marker = "sys_platform != 'darwin'" },
+    { name = "huggingface-hub", version = "1.30.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'darwin'" },
     { name = "humming-kernels", extra = ["cu13"], marker = "sys_platform != 'darwin'" },
     { name = "ijson", marker = "sys_platform != 'darwin'" },
     { name = "jsonschema", marker = "sys_platform != 'darwin'" },
@@ -3835,10 +3861,10 @@ dependencies = [
     { name = "watchfiles", marker = "sys_platform != 'darwin'" },
     { name = "xgrammar", marker = "(platform_machine == 'aarch64' and sys_platform != 'darwin') or (platform_machine == 'arm64' and sys_platform != 'darwin') or (platform_machine == 'ppc64le' and sys_platform != 'darwin') or (platform_machine == 's390x' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform != 'darwin')" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/97/fa/a3cb3a7baf9f1f1027580b0542e42f655b0f577b8a8d39878c1c0a73e6a9/vllm-0.27.1.tar.gz", hash = "sha256:eec2d54d137ac1e59cb4c39226dfee1943eefc8f4788f5821d7300d6acbdb646", size = 39257426, upload-time = "2026-08-11T10:59:48.04Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/22/39/096d88cfe34ee43a8b6e7c957d1ad6338fa5256474d6ff9a32008c6f8cd6/vllm-0.28.0.tar.gz", hash = "sha256:ac96dd0ec5be9c13f2aa4bfb50498c4727110406f6e4980b58a4097e4d18634a", size = 39991441, upload-time = "2026-08-26T10:08:30.56Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/ff/be/6d309884d11ef02966c06d633e156d2070b986df6fccef25041c4b9510e2/vllm-0.27.1-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:74c1a04548c9016ace8f6c03592edf62517152a0e833f46282556ffbe0796254", size = 307180998, upload-time = "2026-08-11T11:00:37.201Z" },
-    { url = "https://files.pythonhosted.org/packages/8f/73/20ad0cc655a5739d95342075c9afab1a25338d4b584fc71548c0b362404b/vllm-0.27.1-cp38-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:98e9fc2a1ed8549a733c9d1b242e2002b82367da9e29e37801761438cb3a2670", size = 312887766, upload-time = "2026-08-11T11:00:11.706Z" },
+    { url = "https://files.pythonhosted.org/packages/b4/31/007ac11dc756f7e69600b90b08e39f573eed9aaee6258198f9d9a2b06355/vllm-0.28.0-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:817b8181f7f61b4a62dc1d5d9ab39f2bfb60a6cb86c29879a78a147b85756787", size = 308873410, upload-time = "2026-08-26T10:08:13.738Z" },
+    { url = "https://files.pythonhosted.org/packages/87/d7/97f6ecc2ae883e601e08d7cef87cb54ececeefcfe6b5e12d5d92f8d06d6b/vllm-0.28.0-cp38-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:addb0ffdaafd8155d75e9b3f5ddb3da28fdee9e8a7097ede91f7db2e9e1a3889", size = 314500103, upload-time = "2026-08-26T10:07:25.091Z" },
 ]
 
 [[package]]