BlogLLM Benchmark Scores as Published, Side by Side
MATH-500: Results as Published
Every result the catalog holds for MATH-500: 26 models from 8 developers, read from 5 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Science and MathMaintainer OpenAI (subset of the MATH dataset)Introduced 2023Data version 2026103001, as of 2026-10-30
MATH-500 is a 500-problem subset of the MATH competition-math test set, released with OpenAI's PRM800K work so that models can be compared on one fixed slice. The score is the share of problems whose final answer is correct.
MATH-500
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-4.1OpenAI, released 2025-04-14 | 92.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| GPT-4.5OpenAI, released 2025-02-27 | 90.7% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15 | - |
| o3-miniOpenAI, released 2025-01-31 | 97.9% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15 | Effort: high |
| o1OpenAI, released 2024-12-17 | 96.4% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15; printed on 1 more page | - |
| o1-miniOpenAI, released 2024-09-12 | 90% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | Model as printed: OpenAI o1-mini |
| 90% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | Model as printed: o1-mini | |
| GPT-4oOpenAI, released 2024-05-13 | 74.6% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15; printed on 1 more page | Model as printed: GPT-4o 0513 |
| 74.6% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | Model as printed: GPT-4o-0513 | |
| Claude Opus 4Anthropic, released 2025-05-22 | 94.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Claude Sonnet 4Anthropic, released 2025-05-22 | 94% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Claude 3.7 SonnetAnthropic, released 2025-02-24 | 96.2% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15; printed on 2 more pages | Effort: extended thinking, 64K |
| 82.2% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15 | Effort: none | |
| Claude 3.5 SonnetAnthropic, released 2024-10-22 | 78% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15; printed on 1 more page | - |
| Gemini 2.5 FlashGoogle, released 2025-05-20 | 95.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| Llama 4 MaverickMeta, released 2025-04-05 | 90% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| DeepSeek-V3-0324DeepSeek, released 2025-03-24 | 94% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 2 more pages | - |
| DeepSeek-R1DeepSeek, released 2025-01-20 | 97.3% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15; printed on 1 more page | - |
| DeepSeek-R1-Distill-Llama-70BDeepSeek, released 2025-01-20 | 94.5% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-R1-Distill-Llama-8BDeepSeek, released 2025-01-20 | 89.1% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-R1-Distill-Qwen-1.5BDeepSeek, released 2025-01-20 | 83.9% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-R1-Distill-Qwen-14BDeepSeek, released 2025-01-20 | 93.9% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-R1-Distill-Qwen-32BDeepSeek, released 2025-01-20 | 94.3% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-R1-Distill-Qwen-7BDeepSeek, released 2025-01-20 | 92.8% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-V3DeepSeek, released 2024-12-26 | 90.2% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 1 more page | - |
| Qwen3-235B-A22BAlibaba, released 2025-04-29 | 91.2% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: non-thinking |
| QwQ-32B-PreviewAlibaba, released 2024-11-27 | 90.6% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| Qwen-MaxAlibaba | 82.6% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15 | - |
| Kimi K2 InstructMoonshot AI, released 2025-07-11 | 97.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| Mistral Medium 3Mistral AI, released 2025-05-07 | 91% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.