Model comparison
DeepSeek-R1 vs Mistral Large 3
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 39.1 on the Noometry Index. Mistral Large 3 costs 2.4× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 20 benchmarks with published results for both. DeepSeek-R1 scores higher in 6 categories and Mistral Large 3 in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-R1 leads 46.3 to 34.4.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 69.4% for DeepSeek-R1 and 50.9% for Mistral Large 3.
- Mistral Large 3 is cheaper at $0.25 / $0.75 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- Mistral Large 3 accepts more context: 262K tokens versus 164K.
- Mistral Large 3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Mistral Large 3 | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 42.3 | 39.1 |
| Released | 2025-01-20 | 2025-12-02 |
| Weights | Proprietary | Open |
| Context window | 164K | 262K |
| Max output | 64K | 8K |
| Input $ / M tokens | $0.50 | $0.25 |
| Output $ / M tokens | $2.15 | $0.75 |
| Results tracked | 52 | 24 |
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 3: 34.4 (#237)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Coding | 1427 | 1448 |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1230 |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Not comparable
DeepSeek-R1: 30.7 (#75), Mistral Large 3: —
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Mistral Large 3: 15.2 (#319)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| Kagi LLM Benchmark | 69.4% | 50.9% |
| LMArena Hard Prompts | 1416 | 1429 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| NYT Connections (extended) | — | 7.5% |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Thematic Generalization | — | 23% |
| LiveBench Reasoning | 83.2% | — |
| LiveBench Data Analysis | 69.8% | — |
| Epoch Capabilities Index | 141.29 | — |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Mistral Large 3: 38.7 (#129)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Math | 1400 | 1414 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Mistral Large 3: 36.0 (#177)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| Vectara Hallucination Rate | 11.3% | 14.5% |
| LMArena Expert | 1394 | 1421 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Mistral Large 3: 38.2 (#66)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Vision | — | 1221 |
Multilingual Too close to call
DeepSeek-R1: 52.4 (#85), Mistral Large 3: 52.5 (#84)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Non-English | 1412 | 1413 |
| LMArena Chinese | 1442 | 1447 |
| LMArena French | 1417 | 1455 |
| LMArena German | 1404 | 1437 |
| LMArena Japanese | 1391 | 1394 |
| LMArena Korean | 1360 | 1384 |
| LMArena Russian | 1423 | 1411 |
| LMArena Spanish | 1411 | 1440 |
Instruction Following Mistral Large 3 leads
DeepSeek-R1: 72.0 (#143), Mistral Large 3: 74.0 (#108)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Instruction Following | 1382 | 1403 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Mistral Large 3: 43.1 (#105)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Longer Query | 1391 | 1413 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Mistral Large 3: 60.0 (#101)
| Benchmark | DeepSeek-R1 | Mistral Large 3 |
|---|---|---|
| LMArena Text | 1428 | 1428 |
| LMArena Creative Writing | 1405 | 1386 |
| EQ-Bench Creative Writing | 1500 | 1412 |
| LMArena Multi-Turn | 1405 | 1429 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Mistral Large 3?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 39.1 on the Noometry Index. Mistral Large 3 costs 2.4× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or Mistral Large 3?
Mistral Large 3 is cheaper. It lists at $0.25 per million input tokens and $0.75 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or Mistral Large 3 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Mistral Large 3 does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-R1 and Mistral Large 3 share?
20 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Mistral Large 3 has 24.