QwQ-32B
A model focused on advancing AI reasoning capabilities, particularly excelling in mathematics and programming. Features deep introspection and self-questioning abilities while having some limitations in language mixing and recursive/endless reasoning patterns.
Benchmark results
| Benchmark | Score | Tags | Source |
|---|---|---|---|
| AIME 2024 | 79.5% | self-reported llm-stats | link → |
| BFCL | 66.4% | self-reported llm-stats | link → |
| GPQA | 65.2% | self-reported llm-stats | link → |
| IFEval | 83.9% | self-reported llm-stats | link → |
| LiveBench | 73.1% | self-reported llm-stats | link → |
| LiveCodeBench | 63.4% | self-reported llm-stats | link → |
| MATH-500 | 90.6% | self-reported llm-stats | link → |