TIR-Bench
TIR-Bench - tool-integrated reasoning benchmark for visual mathematical problem solving. Reports with CI / without CI scores.
Category
Vision & Language
Max Score
—
Score Type
percent
Active
Yes
Model Scores
Scores for TIR-Bench
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.5-4B |
40.62%
|
01.02.2026 | Verified |
| Qwen3.5-122B-A10B |
81.25%
|
01.02.2026 | Verified |
| Qwen3.5-27B |
100.00%
|
01.02.2026 | Unverified |
| Claude-Sonnet-4.5 |
8.52%
|
01.02.2026 | Unverified |
| Qwen3.5-35B-A3B |
87.78%
|
01.02.2026 | Unverified |
| GPT-5-mini 2025-08-07 |
24.60
|
01.02.2026 | Unverified |
| Qwen3-235B-A22B |
14.77%
|
01.02.2026 | Unverified |
| Qwen3-VL-235B-A22B |
14.77%
|
01.02.2026 | Verified |