DeepSeekDeepSeek R1
Best configuration ranks #311 of 610 on the BenchLeader Index at 48.9 ±4.0. Last measured 2 Sept 2026. Released 20 Jan 2025.
- Blended price
- $0.315/M
- $0.280 in · $0.420 out
- Output speed
- 19 tok/s
- First answer
- 1.19 s
- Context
- 131k
- Overall index49
- Reasoning48
- Coding54
- Agents & tools47
- Maths52
- Knowledge46
- Instruction following42
- Human preference61
- Long context44
- Composite46
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | default | Source |
|---|---|---|
| GPQA Diamond | 83.4%#90 | Epoch AI Benchmarking Hub |
| SimpleBench | 40.8%#59 | SimpleBench |
| LMArena Hard Prompts | 1433#114 | LMArena |
| GPQA Diamond (AA)not in index | 81.3%#175 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 15.8%#190 | Artificial Analysis |
| Kagi LLM Benchmark | 69.4%#29 | Kagi LLM Benchmark |
| ARC-AGI-1 | 21.2%#156 | ARC Prize |
| ARC-AGI-2 | 1.3%#153 | ARC Prize |
Coding
| Benchmark | default | Source |
|---|---|---|
| SciCode | 35.7%#130 | SciCode |
| WeirdML | 41.6%#94 | WeirdML |
| LMArena Coding | 1464#109 | LMArena |
| SciCode (AA)not in index | 38.3%#125 | Artificial Analysis |
| LiveCodeBench | 70.2%#86 | Vals AI |
| Aider Polyglot | 74.2%#6 | Aider polyglot leaderboard |
Agents & tools
| Benchmark | default | Source |
|---|---|---|
| Terminal-Bench Hard | 15.9%#184 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 36.5%#223 | Artificial Analysis |
Maths
| Benchmark | default | Source |
|---|---|---|
| OTIS Mock AIME | 87.8%#67 | Epoch AI Benchmarking Hub |
| MATH Level 5 | 96.6%#9 | Epoch AI Benchmarking Hub |
| AIME (Vals) | 74.0%#52 | Vals AI |
| MGSM | 92.3%#20 | Vals AI |
| IMO 2025 | 6.8%#6 | MathArena |
| MathArena Apex | 1.0%#35 | MathArena |
Knowledge
| Benchmark | default | Source |
|---|---|---|
| AA-Omniscience | -27.4#209 | Artificial Analysis |
| MMLU-Pro | 83.2%#72 | Vals AI |
| LegalBench | 67.3%#122 | Vals AI |
| CorpFin | 54.1%#93 | Vals AI |
| TaxEval | 72.3%#66 | Vals AI |
| MedQA | 90.8%#43 | Vals AI |
| PRBench Finance | 32.7%#31 | Scale AI SEAL |
| PRBench Legal | 36.6%#28 | Scale AI SEAL |
| MultiNRC | 27.6%#27 | Scale AI SEAL |
Instruction following
| Benchmark | default | Source |
|---|---|---|
| IFBench | 39.6%#255 | Artificial Analysis |
Human preference
| Benchmark | default | Source |
|---|---|---|
| LMArena Text | 1421#106 | LMArena |
Long context
| Benchmark | default | Source |
|---|---|---|
| Fiction.LiveBench 120k | 33.3%#31 | Fiction.live |
| AA-LCR | 57.7%#227 | Artificial Analysis |
Composite
| Benchmark | default | Source |
|---|---|---|
| Epoch Capabilities Indexnot in index | 141.3#92 | Epoch AI Benchmarking Hub |
| AA Intelligence Index | 13.1#251 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| NovitaAI | 19 tok/s | 1.19 s | $0.700 | $2.50 | 64k | fp8 |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.350 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0002 | $0.0003 | 17.0 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0036 | $0.0042 | 32.8 s |
| Code edit | 6,000 / 1,500 | $0.0023 | $0.0026 | 1.3 min |
| Agentic coding session | 60,000 / 4,000 | $0.018 | $0.022 | 3.5 min |
| Structured extraction | 2,000 / 200 | $0.0006 | $0.0007 | 11.7 s |
See also
Data as of 9 Sept 2026. Compare with another model.