diff --git a/SRC/ctgex2.f b/SRC/ctgex2.f
index 99b18109b..c289dd674 100644
--- a/SRC/ctgex2.f
+++ b/SRC/ctgex2.f
@@ -206,8 +206,8 @@ SUBROUTINE CTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
       COMPLEX            CZERO, CONE
       PARAMETER          ( CZERO = ( 0.0E+0, 0.0E+0 ),
      $                   CONE = ( 1.0E+0, 0.0E+0 ) )
-      REAL               TWENTY
-      PARAMETER          ( TWENTY = 2.0E+1 )
+      REAL               HUNDRED
+      PARAMETER          ( HUNDRED = 1.0E+2 )
       INTEGER            LDST
       PARAMETER          ( LDST = 2 )
       LOGICAL            WANDS
@@ -273,9 +273,11 @@ SUBROUTINE CTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     on 04/01/10.
 *     "Bug" reported by Ondra Kamenik, confirmed by Julie Langou, fixed by
 *     Jim Demmel and Guillaume Revy. See forum post 1783.
+*     Then from TWENTY to HUNDRED on 03/20/26.
+*     "Bug" reported by Heiko Weichelt and Bobby Cheng from MathWorks.
 *
-      THRESHA = MAX( TWENTY*EPS*SA, SMLNUM )
-      THRESHB = MAX( TWENTY*EPS*SB, SMLNUM )
+       THRESHA = MAX( HUNDRED*EPS*SA, SMLNUM )
+       THRESHB = MAX( HUNDRED*EPS*SB, SMLNUM )
 *
 *     Compute unitary QL and RQ that swap 1-by-1 and 1-by-1 blocks
 *     using Givens rotations and perform the swap tentatively.
diff --git a/SRC/dtgex2.f b/SRC/dtgex2.f
index 8c01a4f5e..1fbbb118f 100644
--- a/SRC/dtgex2.f
+++ b/SRC/dtgex2.f
@@ -238,8 +238,8 @@ SUBROUTINE DTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     .. Parameters ..
       DOUBLE PRECISION   ZERO, ONE
       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
-      DOUBLE PRECISION   TWENTY
-      PARAMETER          ( TWENTY = 2.0D+01 )
+      DOUBLE PRECISION   HUNDRED
+      PARAMETER          ( HUNDRED = 1.0D+02 )
       INTEGER            LDST
       PARAMETER          ( LDST = 4 )
       LOGICAL            WANDS
@@ -322,9 +322,11 @@ SUBROUTINE DTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     on 04/01/10.
 *     "Bug" reported by Ondra Kamenik, confirmed by Julie Langou, fixed by
 *     Jim Demmel and Guillaume Revy. See forum post 1783.
+*     Then from TWENTY to HUNDRED on 03/20/26.
+*     "Bug" reported by Heiko Weichelt and Bobby Cheng from MathWorks.
 *
-      THRESHA = MAX( TWENTY*EPS*DNORMA, SMLNUM )
-      THRESHB = MAX( TWENTY*EPS*DNORMB, SMLNUM )
+       THRESHA = MAX( HUNDRED*EPS*DNORMA, SMLNUM )
+       THRESHB = MAX( HUNDRED*EPS*DNORMB, SMLNUM )
 *
       IF( M.EQ.2 ) THEN
 *
diff --git a/SRC/stgex2.f b/SRC/stgex2.f
index d6a11841d..ca5e1c577 100644
--- a/SRC/stgex2.f
+++ b/SRC/stgex2.f
@@ -238,8 +238,8 @@ SUBROUTINE STGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     .. Parameters ..
       REAL               ZERO, ONE
       PARAMETER          ( ZERO = 0.0E+0, ONE = 1.0E+0 )
-      REAL               TWENTY
-      PARAMETER          ( TWENTY = 2.0E+01 )
+      REAL               HUNDRED
+      PARAMETER          ( HUNDRED = 1.0E+02 )
       INTEGER            LDST
       PARAMETER          ( LDST = 4 )
       LOGICAL            WANDS
@@ -323,9 +323,11 @@ SUBROUTINE STGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     on 04/01/10.
 *     "Bug" reported by Ondra Kamenik, confirmed by Julie Langou, fixed by
 *     Jim Demmel and Guillaume Revy. See forum post 1783.
+*     Then from TWENTY to HUNDRED on 03/20/26.
+*     "Bug" reported by Heiko Weichelt and Bobby Cheng from MathWorks.
 *
-      THRESHA = MAX( TWENTY*EPS*DNORMA, SMLNUM )
-      THRESHB = MAX( TWENTY*EPS*DNORMB, SMLNUM )
+       THRESHA = MAX( HUNDRED*EPS*DNORMA, SMLNUM )
+       THRESHB = MAX( HUNDRED*EPS*DNORMB, SMLNUM )
 *
       IF( M.EQ.2 ) THEN
 *
diff --git a/SRC/ztgex2.f b/SRC/ztgex2.f
index 5f97848d4..6fa209791 100644
--- a/SRC/ztgex2.f
+++ b/SRC/ztgex2.f
@@ -206,8 +206,8 @@ SUBROUTINE ZTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
       COMPLEX*16         CZERO, CONE
       PARAMETER          ( CZERO = ( 0.0D+0, 0.0D+0 ),
      $                   CONE = ( 1.0D+0, 0.0D+0 ) )
-      DOUBLE PRECISION   TWENTY
-      PARAMETER          ( TWENTY = 2.0D+1 )
+      DOUBLE PRECISION   HUNDRED
+      PARAMETER          ( HUNDRED = 1.0D+2 )
       INTEGER            LDST
       PARAMETER          ( LDST = 2 )
       LOGICAL            WANDS
@@ -273,9 +273,11 @@ SUBROUTINE ZTGEX2( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z,
 *     on 04/01/10.
 *     "Bug" reported by Ondra Kamenik, confirmed by Julie Langou, fixed by
 *     Jim Demmel and Guillaume Revy. See forum post 1783.
+*     Then from TWENTY to HUNDRED on 03/20/26.
+*     "Bug" reported by Heiko Weichelt and Bobby Cheng from MathWorks.
 *
-      THRESHA = MAX( TWENTY*EPS*SA, SMLNUM )
-      THRESHB = MAX( TWENTY*EPS*SB, SMLNUM )
+       THRESHA = MAX( HUNDRED*EPS*SA, SMLNUM )
+       THRESHB = MAX( HUNDRED*EPS*SB, SMLNUM )
 *
 *     Compute unitary QL and RQ that swap 1-by-1 and 1-by-1 blocks
 *     using Givens rotations and perform the swap tentatively.