Model comparison
DeepSeek-R1 vs o1
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 40.9 on the Noometry Index.
Last verified . 36 shared benchmarks.
DeepSeekSummary
- They share 36 benchmarks with published results for both. DeepSeek-R1 scores higher in 6 categories and o1 in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where o1 leads 27.9 to 18.6.
- The biggest single-benchmark swing is LiveBench Language: 48.5% for DeepSeek-R1 and 65.4% for o1.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $15 / $60 for o1.
- o1 accepts more context: 200K tokens versus 164K.
Side by side
| DeepSeek-R1 | o1 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.3 | 40.9 |
| Released | 2025-01-20 | 2024-09-12 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 200K |
| Max output | 64K | 100K |
| Input $ / M tokens | $0.50 | $15 |
| Output $ / M tokens | $2.15 | $60 |
| Results tracked | 52 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-R1: 46.3 (#68), o1: 46.1 (#70)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| Aider Polyglot | 71.4% | 61.7% |
| WeirdML | 41.6% | 47.6% |
| LiveBench Coding | 66.7% | 69.7% |
| LMArena Coding | 1427 | 1367 |
| SciCode | 35.7% | — |
| CadEval | — | 56% |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
| HumanEval+ | — | 89% |
| MBPP+ | — | 80.2% |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), o1: 24.6 (#117)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| METR Time Horizons | 53.8% | 51.1% |
| Cybench | — | 10% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
Reasoning o1 leads
DeepSeek-R1: 18.6 (#278), o1: 27.9 (#111)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| SimpleBench | 40.8% | 41.7% |
| ARC-AGI-1 | 21.2% | 30.7% |
| LiveBench Reasoning | 83.2% | 91.6% |
| LMArena Hard Prompts | 1416 | 1371 |
| LiveBench Data Analysis | 69.8% | 65.5% |
| Epoch Capabilities Index | 141.29 | 141.91 |
| LiveBench | 71.6% | 75.7% |
| ARC-AGI-2 | 1.3% | — |
| Kagi LLM Benchmark | 69.4% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 15% |
| EnigmaEval | — | 5.7% |
| DTBench | — | 74.7% |
| LMCA | — | 22.3% |
| ForecastBench | 60 | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), o1: 36.1 (#175)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 73.3% |
| LiveBench Math | 80.7% | 80.3% |
| LMArena Math | 1400 | 1388 |
| MATH Level 5 | 96.6% | 94.7% |
| FrontierMath (Tiers 1-3) | — | 14.7% |
| Omni-MATH | 42.4% | — |
| FrontierMath (Feb 2025 set) | — | 9.3% |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), o1: 41.5 (#110)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| GPQA Diamond | 76.3% | 76.8% |
| Confabulations | 12.7% | 11.7% |
| LMArena Expert | 1394 | 1361 |
| Humanity's Last Exam | — | 8% |
| SimpleQA Verified | — | 41.1% |
| MMLU-Pro | 79.3% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, o1: 34.2 (#93)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| LMArena Vision | — | 1168 |
| GeoBench | — | 80% |
| VPCT | — | 37% |
| SpatialViz-Bench | — | 41.4% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), o1: 48.6 (#142)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| LMArena Non-English | 1412 | 1358 |
| LMArena Chinese | 1442 | 1394 |
| LMArena French | 1417 | 1344 |
| LMArena German | 1404 | 1337 |
| LMArena Japanese | 1391 | 1346 |
| LMArena Korean | 1360 | 1396 |
| LMArena Russian | 1423 | 1356 |
| LMArena Spanish | 1411 | 1345 |
Instruction Following o1 leads
DeepSeek-R1: 72.0 (#143), o1: 74.8 (#86)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 81.5% |
| LMArena Instruction Following | 1382 | 1367 |
| IFEval | 78.4% | — |
Long Context o1 leads
DeepSeek-R1: 45.4 (#36), o1: 50.3 (#9)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| Fiction.LiveBench | 75% | 83.3% |
| LMArena Longer Query | 1391 | 1378 |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), o1: 55.6 (#144)
| Benchmark | DeepSeek-R1 | o1 |
|---|---|---|
| LMArena Text | 1428 | 1366 |
| LMArena Creative Writing | 1405 | 1348 |
| Short-Story Creative Writing | 83% | 70.2% |
| LMArena Multi-Turn | 1405 | 1369 |
| LiveBench Language | 48.5% | 65.4% |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
Frequently asked questions
Is DeepSeek-R1 better than o1?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 40.9 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or o1?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; o1 lists at $15 and $60.
Is DeepSeek-R1 or o1 better for coding?
They score almost the same on coding (46.3 vs 46.1); test both on your own repository before choosing.
Which has the bigger context window?
o1 does, with 200K tokens against 164K.
How many benchmarks do DeepSeek-R1 and o1 share?
36 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and o1 has 52.