Benchmark
CritPT
CritPT is a challenging reasoning benchmark reported by Qwen for evaluating frontier mathematical and critical problem-solving capability.
- Modality
- text
- Categories
- math, reasoning
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 5
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| GLM-5.2 | Z.ai | 0.167 | 2026-06-16 | evidence |
| Inkling-Small | Thinking Machines Lab | 0.083 | 2026-07-30 | evidence |
| Nemotron 3 Ultra (550B A55B) | NVIDIA | 0.031 | 2026-06-04 | evidence |
| Qwen3.7 Max | Qwen | 0.114 | 2026-05-19 | evidence |
| Qwen3.7-Plus | Qwen | 0.06 | 2026-05-31 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.