DeepSeekDeepSeek R1
DeepSeek R1 is a DeepSeek open-weights reasoning model, released 20 Jan 2025. Its best configuration ranks #473 of 785 on the BenchLeader Index at 46.8 ±3.6, in the lower half. The ± is the point: 289 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in maths (57) and lowest in agents & tools (36). At $2.50 per million tokens blended it is pricier than most ranked models. Output speed of 144 tokens per second puts it in the fastest quarter, with a first answer in 17.4 s. It has been superseded by DeepSeek R1 0528, which is the version to compare against today. Last measured 22 Oct 2026.
- Blended price
- $2.50/M
- $2.00 in · $4.00 out
- Output speed
- 144 tok/s
- measured by Artificial Analysis
- First answer
- 17 s
- Context
- 66k
- Full answer
- 21 s
- median, reasoning included
- Cost per run
- $0.262
- one full Intelligence Index run
- Released
- 20 Jan 2025
- Overall index47
- Reasoning43
- Coding50
- Agents & tools36
- Maths57
- Knowledge48
- Instruction following42
- Human preference57
- Long context44
- Composite42
Versions
DeepSeek has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| DeepSeek R1 0528 | 28 May 2025 | 49.5 | #392 |
| DeepSeek R1this page | 20 Jan 2025 | 46.8 | #473 |
How its index has moved
8 Sept 2026 to 21 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. 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 | not stated | Source | Trend |
|---|---|---|---|
| GPQA Diamond | 71.7%#163 | Epoch AI Benchmarking Hub | |
| SimpleBench | 30.9%#68 | SimpleBench | |
| Arena Hard Prompts | 1419#151 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 70.8%#292 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 8.5%#350 | Artificial Analysis | |
| Kagi LLM Benchmark | 69.4%#29 | Kagi LLM Benchmark | |
| ARC-AGI-1 | 15.8%#227 | ARC Prize | |
| ARC-AGI-2 | 1.3%#218 | ARC Prize | |
| CritPt | 0.6%#254 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| SciCode | 38.3%#178 | SciCode | |
| WeirdML | 36.5%#123 | WeirdML | |
| Arena Coding | 1444#152 | Arena (LMArena) | |
| SciCode (AA)not in index | 38.3%#188 | Artificial Analysis | |
| LiveCodeBench | 70.2%#86 | Vals AI | |
| Aider Polyglot | 56.9%#18 | Aider polyglot leaderboard | |
| AlgoTunenot in index | 1.7#7 | AlgoTune | |
| Terminal-Bench 4.0 (AA)not in index | 0.0%#152 | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 19.1%#177 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Terminal-Bench Hard | 6.1%#266 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 11.4%#376 | Artificial Analysis | |
| τ³-Banking (AA) | 6.4%#167 | Artificial Analysis | |
| BALROGnot in index | 34.9%#12 | BALROG | |
| METR Time Horizonsnot in index | 51.9%#25 | METR | |
| AutomationBenchnot in index | 0.2%#215 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 2.4%#188 | Artificial Analysis |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| OTIS Mock AIME | 53.3%#178 | Epoch AI Benchmarking Hub | |
| MATH Level 5 | 93.0%#18 | Epoch AI Benchmarking Hub | |
| AIME (Vals) | 74.0%#52 | Vals AI | |
| MGSM | 92.3%#20 | Vals AI | |
| MATH 500not in index | 92.2%#18 | Vals AI | |
| Arena Maths | 1412#130 | Arena (LMArena) |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -32.2#289 | Artificial Analysis | |
| MMLU-Pro | 83.2%#71 | Vals AI | |
| LegalBench | 67.3%#128 | Vals AI | |
| CorpFin | 54.1%#93 | Vals AI | |
| TaxEval | 72.3%#65 | Vals AI | |
| MedQA | 90.8%#43 | Vals AI | |
| MultiNRC | 24.3%#30 | Scale AI SEAL | |
| AA-Omniscience: accuracynot in index | 30.5%#180 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 9.7%#428 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFBench | 39.0%#274 | Artificial Analysis | |
| Arena Instruction Followingnot in index | 1397#139 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1398#152 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1374#142 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1410#143 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1399#162 | Arena (LMArena) | |
| Lech Mazur Creative Writingnot in index | 83.0%#9 | Lech Mazur |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Fiction.LiveBench 120k | 33.3%#31 | Fiction.live | |
| AA-LCR | 57.7%#290 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 138.8#118 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 11.4#353 | Artificial Analysis |
Safety & honesty
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 74.4%#1 | Scale AI SEAL | |
| MASKnot in index | 57.3%#40 | Scale AI SEAL |
Price history
The price above, per 1M tokens, every day it changed.
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. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0020 | – | 19.5 s |
| Summarize a 30-page report | 12,000 / 600 | $0.026 | – | 21.6 s |
| Code edit | 6,000 / 1,500 | $0.018 | – | 27.8 s |
| Agentic coding session | 60,000 / 4,000 | $0.136 | – | 45.1 s |
| Structured extraction | 2,000 / 200 | $0.0048 | – | 18.8 s |
More from DeepSeek
See also
Questions people ask
- What are DeepSeek R1's benchmark scores?
- As of 24 Oct 2026, DeepSeek R1 scores 46.8 ±3.6 on the BenchLeader Index, strongest in maths (57) and weakest in agents & tools (36). That rests on 31 independent benchmark results, and every score on this page links to its publisher.
- How does DeepSeek R1 rank?
- DeepSeek R1 is #473 of 785 ranked configurations on 24 Oct 2026. 289 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does DeepSeek R1 cost per 1M tokens?
- As of 24 Oct 2026, DeepSeek R1 costs $2.00 per 1M input tokens and $4.00 per 1M output tokens, $2.50 blended (three input tokens per output token). The list price last changed on 9 Sept 2026.
- When was DeepSeek R1 released?
- DeepSeek released DeepSeek R1 on 20 Jan 2025.
Data as of 24 Oct 2026. Compare with another model.
Cite as: BenchLeader, “DeepSeek R1: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/deepseek-r1.