[BOT ISSUE] Together: update stale Qwen/Qwen3.7-Max pricing ($1.25/$3.75 → $2.00/$6.00)

[github-actions]

Summary

The Qwen/Qwen3.7-Max model in model_list.json has stale pricing. The current catalog entry shows $1.25 input / $3.75 output per million tokens, but Together AI's pricing page now shows $2.00 input / $6.00 output. The cached input price has also changed from $0.13 to $0.25.

Gap details

Field	Current (local)	Together pricing page	Together docs page
input_cost_per_mil_tokens	1.25	2.00	2.50
output_cost_per_mil_tokens	3.75	6.00	7.50
input_cache_read_cost_per_mil_tokens	0.13	0.25	0.50

The model was originally added via issues #622 and #665 (both closed) with pricing sourced from the Together pricing page at that time ($1.25/$3.75). The pricing page has since been updated. The Together docs page (serverless-models) shows higher values ($2.50/$7.50) that disagree with the pricing page — the pricing page (together.ai/pricing) is used here as the primary authoritative source, consistent with how the model was originally added.

Verification checklist

Check	Status	Detail
Cross-source	✅	Both Together pricing page ($2.00/$6.00) and Together serverless models page ($2.50/$7.50) show prices significantly above the local catalog's $1.25/$3.75. Sources disagree on exact amount; pricing page used as primary
Already in catalog	✅	Qwen/Qwen3.7-Max exists at available_providers: ["together"] with $1.25/$3.75 pricing
ID format	✅	Model ID unchanged
Duplicate check	✅	Issues #622 and #665 covered initial addition (closed/completed). No open issue covers this pricing update

Verification notes

Field	Value	Source
input_cost_per_mil_tokens: 2.00	Verified	Together pricing page
output_cost_per_mil_tokens: 6.00	Verified	Together pricing page
input_cache_read_cost_per_mil_tokens: 0.25	Verified	Together pricing page
reasoning, max_input_tokens, other fields	Unchanged	No update needed for non-pricing fields

Local files inspected

• packages/proxy/schema/model_list.json — Qwen/Qwen3.7-Max at current pricing $1.25/$3.75/$0.13

Note on source discrepancy

Together's two official pages disagree: the pricing page shows $2.00/$6.00/$0.25 while the serverless-models docs page shows $2.50/$7.50/$0.50. The pricing page is the dedicated pricing reference and is used here. The fixer should verify the live price at time of fix.

{
  "kind": "stale_metadata",
  "provider": "together",
  "models": ["Qwen/Qwen3.7-Max"],
  "status": "active",
  "model_specs": {
    "Qwen/Qwen3.7-Max": {
      "format": "openai",
      "flavor": "chat",
      "input_cost_per_mil_tokens": 2.00,
      "output_cost_per_mil_tokens": 6.00,
      "input_cache_read_cost_per_mil_tokens": 0.25,
      "displayName": "Qwen3.7 Max",
      "reasoning": true,
      "max_input_tokens": 1000000,
      "available_providers": ["together"]
    }
  },
  "source_urls": [
    "https://www.together.ai/pricing",
    "https://docs.together.ai/docs/serverless-models"
  ]
}

[github-actions]

Prepared catalog updates for Qwen/Qwen3.7-Max. Updated models: Qwen/Qwen3.7-Max.

Verification sources: https://www.together.ai/pricing, https://docs.together.ai/docs/serverless-models
Verified metadata applied:

• Qwen/Qwen3.7-Max: format=openai, flavor=chat, displayName=Qwen3.7 Max, max_input_tokens=1000000, max_output_tokens=not provided, pricing(input/output per 1M)=2/6, input_cache_read_cost_per_mil_tokens=0.25, providers=together, reasoning=true

Included in the daily batch PR: #1247