Model comparison
DeepSeek V4 Pro vs Mistral Medium
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 36.3 on the Noometry Index.
Last verified . 30 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 30 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 9 categories and Mistral Medium in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.8 to 28.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.6% for DeepSeek V4 Pro and 32.2% for Mistral Medium.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 262K.
Side by side
| DeepSeek V4 Pro | Mistral Medium | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 54.3 | 36.3 |
| Released | 2026-04-24 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 1M | 262K |
| Max output | 393K | 262K |
| Input $ / M tokens | $0.66 | $1.50 |
| Output $ / M tokens | $1.98 | $7.50 |
| Results tracked | 48 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Mistral Medium: 34.2 (#243)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| FrontierCode | 28.6% | 8% |
| SciCode | 51% | 40.2% |
| WeirdML | 66.2% | 43.7% |
| LMArena Coding | 1470 | 1434 |
| ALE-Bench | 1,403 | 763.98 |
| SWE-bench Verified | 77.6% | — |
| LMArena WebDev | 1582 | — |
Agentic & Tool Use DeepSeek V4 Pro leads
DeepSeek V4 Pro: 32.8 (#58), Mistral Medium: 28.3 (#90)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Berkeley Function Calling Leaderboard | — | 37.7% |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Mistral Medium: 24.0 (#167)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 53.5% | 50% |
| CritPt | 18% | 0% |
| LMArena Hard Prompts | 1461 | 1426 |
| DTBench | 93.9% | 75.5% |
| LMCA | 45.5% | 26.1% |
| Surface Evolver Bench | 40% | 26.9% |
| ARC-AGI-2 | 61.3% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| Chess Puzzles | 47% | — |
| Mystery Game Puzzles | 43% | — |
| Epoch Capabilities Index | 155.31 | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Mistral Medium: 28.1 (#245)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 32.2% |
| ProofBench | 50% | 9% |
| LMArena Math | 1455 | 1408 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| MATH Level 5 | — | 81.6% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Mistral Medium: 25.0 (#265)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| GPQA Diamond | 91.7% | 59.5% |
| Vectara Hallucination Rate | 8.6% | 22.7% |
| LMArena Expert | 1464 | 1408 |
| Humanity's Last Exam | — | 4.5% |
| SimpleQA Verified | 52.9% | — |
Multimodal Not comparable
DeepSeek V4 Pro: —, Mistral Medium: 35.3 (#88)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Mistral Medium: 52.1 (#91)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1439 | 1408 |
| LMArena Chinese | 1486 | 1447 |
| LMArena French | 1472 | 1459 |
| LMArena German | 1458 | 1432 |
| LMArena Japanese | 1445 | 1378 |
| LMArena Korean | 1447 | 1380 |
| LMArena Russian | 1453 | 1411 |
| LMArena Spanish | 1458 | 1433 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), Mistral Medium: 73.7 (#116)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1448 | 1398 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Mistral Medium: 42.9 (#114)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1458 | 1406 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Mistral Medium: 60.0 (#103)
| Benchmark | DeepSeek V4 Pro | Mistral Medium |
|---|---|---|
| LMArena Text | 1451 | 1424 |
| LMArena Creative Writing | 1446 | 1391 |
| LMArena Multi-Turn | 1467 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Mistral Medium?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 36.3 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Mistral Medium?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is DeepSeek V4 Pro or Mistral Medium better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 262K.
How many benchmarks do DeepSeek V4 Pro and Mistral Medium share?
30 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Mistral Medium has 36.