Model comparison
DeepSeek V4 Flash vs Devstral 2
DeepSeek V4 Flash has enough public results to be ranked (#32); Devstral 2 does not yet, so treat this comparison as directional.
Last verified . 1 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 1 benchmark with published results for both. DeepSeek V4 Flash scores higher in 1 category and Devstral 2 in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where DeepSeek V4 Flash leads 47.6 to 33.0.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.40 / $2 for Devstral 2.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 262K.
Side by side
| DeepSeek V4 Flash | Devstral 2 | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 53.6 | 35.8 |
| Released | 2026-04-24 | 2025-12-09 |
| Weights | Open | Open |
| Context window | 1M | 262K |
| Max output | 393K | 262K |
| Input $ / M tokens | $0.15 | $0.40 |
| Output $ / M tokens | $0.60 | $2 |
| Results tracked | 41 | 2 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Flash leads
DeepSeek V4 Flash: 47.6 (#56), Devstral 2: 33.0 (#253)
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| LMArena WebDev | 1581 | 1196 |
| FrontierCode | 18.8% | — |
| SWE-bench Verified (bash only) | — | 53.8% |
| SciCode | 49.9% | — |
| WeirdML | 63% | — |
| LMArena Coding | 1457 | — |
| ALE-Bench | 1,306 | — |
Reasoning Not comparable
DeepSeek V4 Flash: 53.7 (#25), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| ARC-AGI-2 | 61.4% | — |
| SimpleBench | 61.1% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| LMArena Hard Prompts | 1443 | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |
| Epoch Capabilities Index | 154.51 | — |
Math Not comparable
DeepSeek V4 Flash: 60.0 (#31), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
| ProofBench | 56% | — |
| LMArena Math | 1427 | — |
Knowledge Not comparable
DeepSeek V4 Flash: 55.5 (#43), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| GPQA Diamond | 91% | — |
| SimpleQA Verified | 33.6% | — |
| LMArena Expert | 1441 | — |
Multilingual Not comparable
DeepSeek V4 Flash: 53.2 (#71), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| LMArena Non-English | 1420 | — |
| LMArena Chinese | 1467 | — |
| LMArena French | 1438 | — |
| LMArena German | 1418 | — |
| LMArena Japanese | 1403 | — |
| LMArena Korean | 1383 | — |
| LMArena Russian | 1428 | — |
| LMArena Spanish | 1437 | — |
Instruction Following Not comparable
DeepSeek V4 Flash: 75.1 (#76), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| LMArena Instruction Following | 1421 | — |
Long Context Not comparable
DeepSeek V4 Flash: 43.8 (#81), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| LMArena Longer Query | 1435 | — |
Writing & Preference Not comparable
DeepSeek V4 Flash: 64.0 (#59), Devstral 2: —
| Benchmark | DeepSeek V4 Flash | Devstral 2 |
|---|---|---|
| LMArena Text | 1432 | — |
| LMArena Creative Writing | 1403 | — |
| EQ-Bench Creative Writing | 1559 | — |
| LMArena Multi-Turn | 1449 | — |
Frequently asked questions
Is DeepSeek V4 Flash better than Devstral 2?
DeepSeek V4 Flash has enough public results to be ranked (#32); Devstral 2 does not yet, so treat this comparison as directional.
Which is cheaper, DeepSeek V4 Flash or Devstral 2?
DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Devstral 2 lists at $0.40 and $2.
Is DeepSeek V4 Flash or Devstral 2 better for coding?
DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 versus 33.0 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Flash does, with 1M tokens against 262K.
How many benchmarks do DeepSeek V4 Flash and Devstral 2 share?
1 benchmark has published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Devstral 2 has 2.