DeepSeek
availableShows if the model has enough results for an index.DeepSeek-R1
DeepSeek-R1 is a reasoning model from DeepSeek. 18 benchmarks count toward its score, in 6 categories.
IndexOverall score. 50 is the middle.40.3 ±5.3
CoverageShare of the index weight with results.85%
SpeedOutput tokens per second.18/s
Input / 1MUS dollars per 1M input tokens.$0.7
Output / 1MUS dollars per 1M output tokens.$2.5
ContextMaximum tokens in one request.64K
EloLMArena rating and rank.1373 (#168)
50 is the middle of the board. The range shows the doubt in the index.
18,524 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category. 50 is the middle.
50 is the middleResults
18 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result on the index scale. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| MATH level 5 | Math | 93.1% | 44.4 | — | — | Epoch AI |
| MGSM | Multilingual | 92.3% | — | — | 9 Jan 2026 | Vals AI |
| MATH 500 | Math | 92.2% | 44.8 | — | 9 Jan 2026 | Vals AI |
| MMLU Pro | Knowledge | 83.2% | 51.6 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis GPQA Diamond | Knowledge | 81.3% | 51.9 | — | — | Artificial Analysis |
| AIME | Math | 74.0% | 48.2 | — | 16 Apr 2026 | Vals AI |
| GPQA diamond | Knowledge | 71.7% | 44.5 | — | — | Epoch AI |
| LiveCodeBench | Coding | 70.2% | 48.2 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis Long Context Reasoning | Reasoning | 55.7% | 46.8 | — | — | Artificial Analysis |
| OTIS Mock AIME 2024-2025 | Math | 53.3% | 41.4 | — | — | Epoch AI |
| Artificial Analysis IFBench | Instruction | 39.6% | 29.3 | — | — | Artificial Analysis |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 36.5% | 25.0 | — | — | Victor Barres et al. |
| Artificial Analysis Omniscience Accuracy | Knowledge | 30.5% | 51.6 | — | — | Artificial Analysis |
| Artificial Analysis Humanity's Last Exam | Knowledge | 15.8% | 41.7 | — | — | Artificial Analysis |
| ARC-AGI-1 (semi-private) | Reasoning | 15.8% | 34.7 | — | — | ARC Prize Foundation |
| Terminal-Bench 1.0 | Agentic | 13.8% | 23.2 | — | 12 Jan 2026 | Vals AI |
| Artificial Analysis Intelligence Index | Knowledge | 13.1% | 38.8 | — | — | Artificial Analysis |
| Critical Physics Tasks | Reasoning | 1.4% | 41.3 | — | — | Artificial Analysis |
| ARC-AGI-2 (semi-private) | Reasoning | 1.3% | 40.4 | — | — | ARC Prize Foundation |
18 benchmarks count, from 18 of 19 results. A grey row does not count. Too few models took that benchmark.