Cogito v2.1 (Deep Cogito research blog) 19 Nov 2025
- SWE-Bench Verified (solve rate, Agentless) 42
- GPQA Diamond 77.7
- MMLU Pro 84.7
- Multilingual MMLU 86.2
- Humanity's Last Exam 16.7
- SimpleQA Verified (f1) 30.9
- AIME 2025 89.5
- MATH-500 98.6
Cogito v2.1
Cogito v2.1 places in the top three on MATH-500.
Deep Cogito Released 19 Nov 2025 9
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Deep Cogito Nov 2025 Cogito v2.1
Open weights
No Epoch Capabilities Index score yet
- CritPt 0% #145 /160
- MMLU-Pro 84.7% #33 /151
- AIME 2025 89.5% #52 /143
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMLU-Pro | 84.7% | #33 of 151 | Model card | |
| MMMLU | 86.2% | #25 of 62 | Model card | |
| SimpleQA Verified | 30.9% | #16 of 22 | Model card | |
| Humanity's Last Exam | 16.7% | #114 of 147 | Model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MATH-500 | 98.6% | #3 of 67 | Model card | |
| AIME 2025 | 89.5% | #52 of 143 | Model card | |
| GPQA Diamond | 77.7% | #124 of 284 | Model card | |
| CritPt | 0% | #145 of 160 | Epoch AI |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SWE-bench Verified | 42% | #137 of 150 | Model card |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.