Model comparison
DeepSeek-R1 vs Mistral Large
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 31.9 on the Noometry Index.
Last verified . 42 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 42 benchmarks with published results for both. DeepSeek-R1 scores higher in 9 categories and Mistral Large in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 18.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 8.5% for Mistral Large.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $2 / $6 for Mistral Large.
- DeepSeek-R1 accepts more context: 164K tokens versus 131K.
- Mistral Large has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Mistral Large | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 42.3 | 31.9 |
| Released | 2025-01-20 | 2024-02-26 |
| Weights | Proprietary | Open |
| Context window | 164K | 131K |
| Max output | 64K | 16K |
| Input $ / M tokens | $0.50 | $2 |
| Output $ / M tokens | $2.15 | $6 |
| Results tracked | 52 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Mistral Large: 34.3 (#240)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| SciCode | 35.7% | 36.2% |
| LiveBench Coding | 66.7% | 47.1% |
| LMArena Coding | 1427 | 1277 |
| ALE-Bench | 804.12 | 264.7 |
| Aider Polyglot | 71.4% | — |
| WeirdML | 41.6% | — |
| BigCodeBench Instruct | — | 30% |
| BigCodeBench Complete | — | 38.3% |
| AlgoTune | 1.7 | — |
| HumanEval+ | — | 62.2% |
| MBPP+ | — | 59.5% |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), Mistral Large: 28.6 (#89)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 38.4% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Mistral Large: 15.8 (#310)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| SimpleBench | 40.8% | 22.5% |
| CritPt | 1.1% | 0% |
| LiveBench Reasoning | 83.2% | 43.5% |
| LMArena Hard Prompts | 1416 | 1257 |
| LiveBench Data Analysis | 69.8% | 50.1% |
| Epoch Capabilities Index | 141.29 | 128.52 |
| ForecastBench | 60 | 57.1 |
| LiveBench | 71.6% | 48.4% |
| ARC-AGI-2 | 1.3% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| DTBench | — | 65.1% |
| LMCA | — | 16.7% |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Mistral Large: 18.2 (#291)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 8.5% |
| Omni-MATH | 42.4% | 28.1% |
| LiveBench Math | 80.7% | 42.5% |
| LMArena Math | 1400 | 1262 |
| MATH Level 5 | 96.6% | 50.3% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Mistral Large: 30.1 (#230)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| GPQA Diamond | 76.3% | 51.3% |
| MMLU-Pro | 79.3% | 59.9% |
| Confabulations | 12.7% | 21.4% |
| Vectara Hallucination Rate | 11.3% | 4.5% |
| GPQA (HELM) | 66.6% | 43.5% |
| LMArena Expert | 1394 | 1232 |
| MMLU | — | 80% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Mistral Large: 40.0 (#219)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| LMArena Non-English | 1412 | 1237 |
| LMArena Chinese | 1442 | 1240 |
| LMArena French | 1417 | 1325 |
| LMArena German | 1404 | 1254 |
| LMArena Japanese | 1391 | 1188 |
| LMArena Korean | 1360 | 1202 |
| LMArena Russian | 1423 | 1257 |
| LMArena Spanish | 1411 | 1268 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), Mistral Large: 67.9 (#191)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 67.9% |
| IFEval | 78.4% | 87.7% |
| LMArena Instruction Following | 1382 | 1249 |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Mistral Large: 38.3 (#199)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| LMArena Longer Query | 1391 | 1261 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Mistral Large: 40.7 (#242)
| Benchmark | DeepSeek-R1 | Mistral Large |
|---|---|---|
| LMArena Text | 1428 | 1266 |
| LMArena Creative Writing | 1405 | 1243 |
| Short-Story Creative Writing | 83% | 69% |
| EQ-Bench Creative Writing | 1500 | 985 |
| WildBench | 82.8% | 80.1% |
| LMArena Multi-Turn | 1405 | 1260 |
| LiveBench Language | 48.5% | 39.4% |
Frequently asked questions
Is DeepSeek-R1 better than Mistral Large?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 31.9 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or Mistral Large?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Mistral Large lists at $2 and $6.
Is DeepSeek-R1 or Mistral Large better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 34.3 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-R1 and Mistral Large share?
42 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Mistral Large has 51.