Model comparison
DeepSeek-V2.5 (Sep 2024) vs Mistral Large
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 31.9 on the Noometry Index.
Last verified . 21 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 21 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 6 categories and Mistral Large in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 35.9 to 18.2.
- The biggest single-benchmark swing is BigCodeBench Instruct: 48.6% for DeepSeek-V2.5 (Sep 2024) and 30% for Mistral Large.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Mistral Large | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 37.6 | 31.9 |
| Released | 2024-09-06 | 2024-02-26 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 16K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $6 |
| Results tracked | 22 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Large leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Mistral Large: 34.3 (#240)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| BigCodeBench Instruct | 48.6% | 30% |
| LMArena Coding | 1309 | 1277 |
| BigCodeBench Complete | 53.2% | 38.3% |
| HumanEval+ | 83.5% | 62.2% |
| MBPP+ | 74.1% | 59.5% |
| Aider Polyglot | 17.8% | — |
| SciCode | — | 36.2% |
| LiveBench Coding | — | 47.1% |
| ALE-Bench | — | 264.7 |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Mistral Large: 28.6 (#89)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 38.4% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Mistral Large: 15.8 (#310)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1257 |
| SimpleBench | — | 22.5% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 43.5% |
| DTBench | — | 65.1% |
| LiveBench Data Analysis | — | 50.1% |
| LMCA | — | 16.7% |
| Epoch Capabilities Index | — | 128.52 |
| ForecastBench | — | 57.1 |
| LiveBench | — | 48.4% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Mistral Large: 18.2 (#291)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Math | 1288 | 1262 |
| OTIS Mock AIME 2024-2025 | — | 8.5% |
| Omni-MATH | — | 28.1% |
| LiveBench Math | — | 42.5% |
| MATH Level 5 | — | 50.3% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Mistral Large: 30.1 (#230)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Expert | 1266 | 1232 |
| GPQA Diamond | — | 51.3% |
| MMLU-Pro | — | 59.9% |
| Confabulations | — | 21.4% |
| Vectara Hallucination Rate | — | 4.5% |
| GPQA (HELM) | — | 43.5% |
| MMLU | — | 80% |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Mistral Large: 40.0 (#219)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Non-English | 1273 | 1237 |
| LMArena Chinese | 1318 | 1240 |
| LMArena French | 1289 | 1325 |
| LMArena German | 1258 | 1254 |
| LMArena Japanese | 1228 | 1188 |
| LMArena Korean | 1209 | 1202 |
| LMArena Russian | 1289 | 1257 |
| LMArena Spanish | 1248 | 1268 |
Instruction Following Too close to call
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Mistral Large: 67.9 (#191)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Instruction Following | 1280 | 1249 |
| LiveBench Instruction Following | — | 67.9% |
| IFEval | — | 87.7% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Mistral Large: 38.3 (#199)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Longer Query | 1301 | 1261 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Mistral Large: 40.7 (#242)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mistral Large |
|---|---|---|
| LMArena Text | 1294 | 1266 |
| LMArena Creative Writing | 1285 | 1243 |
| LMArena Multi-Turn | 1297 | 1260 |
| Short-Story Creative Writing | — | 69% |
| EQ-Bench Creative Writing | — | 985 |
| WildBench | — | 80.1% |
| LiveBench Language | — | 39.4% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Mistral Large?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 31.9 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Mistral Large better for coding?
Mistral Large scores higher on coding benchmarks: 34.3 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Mistral Large share?
21 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Mistral Large has 51.