Model comparison
DeepSeek-V3.1 vs Mistral Medium
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 36.3 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 22 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 6 categories and Mistral Medium in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.1 leads 43.7 to 25.0.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 5.5% for DeepSeek-V3.1 and 22.7% for Mistral Medium.
- DeepSeek-V3.1 is cheaper at $0.25 / $0.95 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-V3.1 | Mistral Medium | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 42.8 | 36.3 |
| Released | 2025-08-21 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 8K | 262K |
| Input $ / M tokens | $0.25 | $1.50 |
| Output $ / M tokens | $0.95 | $7.50 |
| Results tracked | 27 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1 leads
DeepSeek-V3.1: 40.3 (#144), Mistral Medium: 34.2 (#243)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| WeirdML | 38.4% | 43.7% |
| LMArena Coding | 1417 | 1434 |
| FrontierCode | — | 8% |
| SciCode | — | 40.2% |
| ALE-Bench | — | 763.98 |
Agentic & Tool Use Not comparable
DeepSeek-V3.1: —, Mistral Medium: 28.3 (#90)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.7% |
Reasoning DeepSeek-V3.1 leads
DeepSeek-V3.1: 27.9 (#110), Mistral Medium: 24.0 (#167)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 50% |
| LMArena Hard Prompts | 1417 | 1426 |
| DTBench | 82.7% | 75.5% |
| LMCA | 24.3% | 26.1% |
| SimpleBench | 40% | — |
| CritPt | — | 0% |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 139.92 | — |
| ForecastBench | 58 | — |
Math DeepSeek-V3.1 leads
DeepSeek-V3.1: 38.9 (#122), Mistral Medium: 28.1 (#245)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Math | 1420 | 1408 |
| OTIS Mock AIME 2024-2025 | — | 32.2% |
| ProofBench | — | 9% |
| MATH Level 5 | — | 81.6% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge DeepSeek-V3.1 leads
DeepSeek-V3.1: 43.7 (#90), Mistral Medium: 25.0 (#265)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| Vectara Hallucination Rate | 5.5% | 22.7% |
| LMArena Expert | 1405 | 1408 |
| GPQA Diamond | — | 59.5% |
| Humanity's Last Exam | — | 4.5% |
Multimodal Not comparable
DeepSeek-V3.1: —, Mistral Medium: 35.3 (#88)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual Too close to call
DeepSeek-V3.1: 51.6 (#106), Mistral Medium: 52.1 (#91)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1400 | 1408 |
| LMArena Chinese | 1469 | 1447 |
| LMArena French | 1447 | 1459 |
| LMArena German | 1411 | 1432 |
| LMArena Japanese | 1378 | 1378 |
| LMArena Korean | 1337 | 1380 |
| LMArena Russian | 1405 | 1411 |
| LMArena Spanish | 1431 | 1433 |
Instruction Following Too close to call
DeepSeek-V3.1: 73.9 (#110), Mistral Medium: 73.7 (#116)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1400 | 1398 |
Long Context Mistral Medium leads
DeepSeek-V3.1: 36.3 (#232), Mistral Medium: 42.9 (#114)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1422 | 1406 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference Too close to call
DeepSeek-V3.1: 60.3 (#98), Mistral Medium: 60.0 (#103)
| Benchmark | DeepSeek-V3.1 | Mistral Medium |
|---|---|---|
| LMArena Text | 1420 | 1424 |
| LMArena Creative Writing | 1401 | 1391 |
| LMArena Multi-Turn | 1408 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| EQ-Bench Creative Writing | 1436 | — |
Frequently asked questions
Is DeepSeek-V3.1 better than Mistral Medium?
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 36.3 on the Noometry Index.
Which is cheaper, DeepSeek-V3.1 or Mistral Medium?
DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is DeepSeek-V3.1 or Mistral Medium better for coding?
DeepSeek-V3.1 scores higher on coding benchmarks: 40.3 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
Mistral Medium does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3.1 and Mistral Medium share?
22 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Mistral Medium has 36.