Model comparison
DeepSeek-V2.5 (Sep 2024) vs Mistral Medium
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 36.0 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 3 categories and Mistral Medium in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Medium leads 60.2 to 50.0.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Mistral Medium | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 37.0 | 36.0 |
| Released | 2024-09-06 | 2023-12-11 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 262K |
| Input $ / M tokens | — | $1.50 |
| Output $ / M tokens | — | $7.50 |
| Results tracked | 22 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Medium leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), Mistral Medium: 33.8 (#241)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Coding | 1310 | 1434 |
| FrontierCode | — | 8% |
| Aider Polyglot | 17.8% | — |
| SciCode | — | 40.2% |
| WeirdML | — | 43.7% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| ALE-Bench | — | 763.98 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Mistral Medium: 28.1 (#84)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.7% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), Mistral Medium: 23.3 (#165)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1426 |
| Kagi LLM Benchmark | — | 50% |
| CritPt | — | 0% |
| DTBench | — | 75.5% |
| LMCA | — | 26.1% |
| Surface Evolver Bench | — | 26.9% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), Mistral Medium: 27.2 (#238)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Math | 1289 | 1409 |
| OTIS Mock AIME 2024-2025 | — | 32.2% |
| ProofBench | — | 9% |
| MATH Level 5 | — | 81.6% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), Mistral Medium: 24.9 (#260)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Expert | 1266 | 1408 |
| GPQA Diamond | — | 59.5% |
| Humanity's Last Exam | — | 4.5% |
| Vectara Hallucination Rate | — | 22.7% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, Mistral Medium: 34.6 (#82)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual Mistral Medium leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), Mistral Medium: 52.3 (#86)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1273 | 1408 |
| LMArena Chinese | 1317 | 1448 |
| LMArena French | 1288 | 1460 |
| LMArena German | 1258 | 1433 |
| LMArena Japanese | 1228 | 1378 |
| LMArena Korean | 1209 | 1382 |
| LMArena Russian | 1289 | 1411 |
| LMArena Spanish | 1248 | 1433 |
Instruction Following Mistral Medium leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), Mistral Medium: 73.9 (#111)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1280 | 1398 |
Long Context Mistral Medium leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), Mistral Medium: 42.9 (#108)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1301 | 1406 |
Writing & Preference Mistral Medium leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), Mistral Medium: 60.2 (#97)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Medium |
|---|---|---|
| LMArena Text | 1294 | 1424 |
| LMArena Creative Writing | 1285 | 1391 |
| LMArena Multi-Turn | 1297 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Mistral Medium?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 36.0 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Mistral Medium better for coding?
Mistral Medium scores higher on coding benchmarks: 33.8 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Mistral Medium share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Mistral Medium has 36.