Model comparison
DeepSeek-V3.2-Exp vs Mistral Medium
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 36.3 on the Noometry Index.
Last verified . 29 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 29 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 8 categories and Mistral Medium in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 25.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 87.8% for DeepSeek-V3.2-Exp and 32.2% for Mistral Medium.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 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-V3.2-Exp | Mistral Medium | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 44.3 | 36.3 |
| Released | 2025-09-29 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 66K | 262K |
| Input $ / M tokens | $0.26 | $1.50 |
| Output $ / M tokens | $0.38 | $7.50 |
| Results tracked | 49 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), Mistral Medium: 34.2 (#243)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| SciCode | 38.9% | 40.2% |
| WeirdML | 39.5% | 43.7% |
| LMArena Coding | 1454 | 1434 |
| FrontierCode | — | 8% |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| ALE-Bench | — | 763.98 |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), Mistral Medium: 28.3 (#90)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | 56.7% | 37.7% |
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Mistral Medium leads
DeepSeek-V3.2-Exp: 22.1 (#208), Mistral Medium: 24.0 (#167)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 50% |
| CritPt | 2.9% | 0% |
| LMArena Hard Prompts | 1434 | 1426 |
| DTBench | 87.7% | 75.5% |
| LMCA | 29.1% | 26.1% |
| ARC-AGI-2 | 4% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| Chess Puzzles | 14% | — |
| Thematic Generalization | 65% | — |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 146.27 | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), Mistral Medium: 28.1 (#245)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 32.2% |
| ProofBench | 8% | 9% |
| LMArena Math | 1435 | 1408 |
| FrontierMath (Feb 2025 set) | 22.1% | 0.3% |
| MathArena Final-Answer Competitions | 57.7% | — |
| MATH Level 5 | — | 81.6% |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), Mistral Medium: 25.0 (#265)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| GPQA Diamond | 83.4% | 59.5% |
| Vectara Hallucination Rate | 5.3% | 22.7% |
| LMArena Expert | 1436 | 1408 |
| Humanity's Last Exam | — | 4.5% |
Multimodal Not comparable
DeepSeek-V3.2-Exp: —, Mistral Medium: 35.3 (#88)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual Too close to call
DeepSeek-V3.2-Exp: 52.2 (#90), Mistral Medium: 52.1 (#91)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1409 | 1408 |
| LMArena Chinese | 1461 | 1447 |
| LMArena French | 1433 | 1459 |
| LMArena German | 1440 | 1432 |
| LMArena Japanese | 1374 | 1378 |
| LMArena Korean | 1371 | 1380 |
| LMArena Russian | 1424 | 1411 |
| LMArena Spanish | 1440 | 1433 |
Instruction Following Too close to call
DeepSeek-V3.2-Exp: 74.5 (#93), Mistral Medium: 73.7 (#116)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1413 | 1398 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), Mistral Medium: 42.9 (#114)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1428 | 1406 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Mistral Medium: 60.0 (#103)
| Benchmark | DeepSeek-V3.2-Exp | Mistral Medium |
|---|---|---|
| LMArena Text | 1425 | 1424 |
| LMArena Creative Writing | 1403 | 1391 |
| LMArena Multi-Turn | 1427 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Mistral Medium?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 36.3 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or Mistral Medium?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is DeepSeek-V3.2-Exp or Mistral Medium better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 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-V3.2-Exp and Mistral Medium share?
29 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Mistral Medium has 36.