Model comparison
DeepSeek V4 Pro vs Magistral Small
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 30.2 on the Noometry Index.
Last verified . 10 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 10 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 4 categories and Magistral Small in 0 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 6.8.
- The biggest single-benchmark swing is ARC-AGI-1: 90.5% for DeepSeek V4 Pro and 5% for Magistral Small.
- Magistral Small is cheaper at $0.50 / $1.50 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 128K.
Side by side
| DeepSeek V4 Pro | Magistral Small | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 54.3 | 30.2 |
| Released | 2026-04-24 | 2025-06-10 |
| Weights | Open | Open |
| Context window | 1M | 128K |
| Max output | 393K | 40K |
| Input $ / M tokens | $0.66 | $0.50 |
| Output $ / M tokens | $1.98 | $1.50 |
| Results tracked | 48 | 10 |
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), Magistral Small: 38.4 (#176)
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| SciCode | 51% | 35.2% |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| LMArena WebDev | 1582 | — |
| WeirdML | 66.2% | — |
| LMArena Coding | 1470 | — |
| ALE-Bench | 1,403 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), Magistral Small: —
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Magistral Small: 6.8 (#350)
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| ARC-AGI-2 | 61.3% | 0% |
| Kagi LLM Benchmark | 53.5% | 6.3% |
| ARC-AGI-1 | 90.5% | 5% |
| CritPt | 18% | 0.3% |
| Chess Puzzles | 47% | 3% |
| DTBench | 93.9% | 61.3% |
| Epoch Capabilities Index | 155.31 | 133.19 |
| NYT Connections (extended) | 91.3% | — |
| LMArena Hard Prompts | 1461 | — |
| Mystery Game Puzzles | 43% | — |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Magistral Small: 26.2 (#261)
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 30% |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |
| LMArena Math | 1455 | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Magistral Small: 30.9 (#223)
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| GPQA Diamond | 91.7% | 56.1% |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| LMArena Expert | 1464 | — |
Multilingual Not comparable
DeepSeek V4 Pro: 54.4 (#45), Magistral Small: —
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| LMArena Non-English | 1439 | — |
| LMArena Chinese | 1486 | — |
| LMArena French | 1472 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1445 | — |
| LMArena Korean | 1447 | — |
| LMArena Russian | 1453 | — |
| LMArena Spanish | 1458 | — |
Instruction Following Not comparable
DeepSeek V4 Pro: 76.1 (#47), Magistral Small: —
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| LMArena Instruction Following | 1448 | — |
Long Context Not comparable
DeepSeek V4 Pro: 45.0 (#51), Magistral Small: —
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| CL-bench Life | 13.5% | — |
| LMArena Longer Query | 1458 | — |
Writing & Preference Not comparable
DeepSeek V4 Pro: 65.5 (#46), Magistral Small: —
| Benchmark | DeepSeek V4 Pro | Magistral Small |
|---|---|---|
| LMArena Text | 1451 | — |
| LMArena Creative Writing | 1446 | — |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LMArena Multi-Turn | 1467 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Magistral Small?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 30.2 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Magistral Small?
Magistral Small is cheaper. It lists at $0.50 per million input tokens and $1.50 per million output tokens; DeepSeek V4 Pro lists at $0.66 and $1.98.
Is DeepSeek V4 Pro or Magistral Small better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 38.4 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 128K.
How many benchmarks do DeepSeek V4 Pro and Magistral Small share?
10 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Magistral Small has 10.