Model comparison
Devstral Small 2505 vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 34.3 on the Noometry Index. Devstral Small 2505 costs 40× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Mistral AISummary
- They share 2 benchmarks with published results for both. Devstral Small 2505 scores higher in 0 categories and Kimi K3 in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K3 leads 63.0 to 19.7.
- The biggest single-benchmark swing is SciCode: 28.8% for Devstral Small 2505 and 59.5% for Kimi K3.
- Devstral Small 2505 is cheaper at $0.10 / $0.30 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 128K.
Side by side
| Devstral Small 2505 | Kimi K3 | |
|---|---|---|
| Provider | Mistral AI | Moonshot AI |
| Noometry Index | 34.3 | 59.5 |
| Released | 2025-05-07 | 2026-07-16 |
| Weights | Open | Open |
| Context window | 128K | 1.05M |
| Max output | 128K | 1.05M |
| Input $ / M tokens | $0.10 | $3 |
| Output $ / M tokens | $0.30 | $15 |
| Results tracked | 4 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Devstral Small 2505: 38.9 (#166), Kimi K3: 61.0 (#10)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| SciCode | 28.8% | 59.5% |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| SWE-bench Verified (bash only) | 56.4% | — |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| WeirdML | — | 82.6% |
| LMArena Coding | — | 1508 |
| ALE-Bench | — | 1,524 |
Agentic & Tool Use Not comparable
Devstral Small 2505: —, Kimi K3: 41.8 (#20)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| APEX-Agents | — | 50.6% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| Vending-Bench 2 | — | 5,165 |
Reasoning Kimi K3 leads
Devstral Small 2505: 19.7 (#252), Kimi K3: 63.0 (#17)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| CritPt | 0% | 23.4% |
| ARC-AGI-2 | — | 60.4% |
| SimpleBench | — | 60.7% |
| Kagi LLM Benchmark | 37.7% | — |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| Chess Puzzles | — | 39% |
| LMArena Hard Prompts | — | 1496 |
| Mystery Game Puzzles | — | 26% |
| DTBench | — | 91.2% |
| LMCA | — | 52.7% |
| Surface Evolver Bench | — | 95% |
| Epoch Capabilities Index | — | 157.45 |
| ForecastBench | — | 61.1 |
Math Not comparable
Devstral Small 2505: —, Kimi K3: 74.2 (#16)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 72.2% |
| FrontierMath Tier 4 | — | 39% |
| MathArena Final-Answer Competitions | — | 87.8% |
| OTIS Mock AIME 2024-2025 | — | 97.2% |
| ProofBench | — | 87% |
| LMArena Math | — | 1491 |
Knowledge Not comparable
Devstral Small 2505: —, Kimi K3: 63.2 (#21)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| GPQA Diamond | — | 93.1% |
| SimpleQA Verified | — | 50.6% |
| LMArena Expert | — | 1521 |
Multimodal Not comparable
Devstral Small 2505: —, Kimi K3: 37.8 (#70)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Not comparable
Devstral Small 2505: —, Kimi K3: 56.3 (#21)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| LMArena Non-English | — | 1466 |
| LMArena Chinese | — | 1529 |
| LMArena French | — | 1491 |
| LMArena German | — | 1488 |
| LMArena Japanese | — | 1487 |
| LMArena Korean | — | 1458 |
| LMArena Russian | — | 1482 |
| LMArena Spanish | — | 1472 |
Instruction Following Not comparable
Devstral Small 2505: —, Kimi K3: 77.7 (#14)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | — | 1483 |
Long Context Not comparable
Devstral Small 2505: —, Kimi K3: 45.8 (#29)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | — | 1494 |
Writing & Preference Not comparable
Devstral Small 2505: —, Kimi K3: 76.6 (#4)
| Benchmark | Devstral Small 2505 | Kimi K3 |
|---|---|---|
| LMArena Text | — | 1476 |
| LMArena Creative Writing | — | 1454 |
| EQ-Bench Creative Writing | — | 2082 |
| EQ-Bench 4 | — | 1339 |
| LMArena Multi-Turn | — | 1488 |
Frequently asked questions
Is Devstral Small 2505 better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 34.3 on the Noometry Index. Devstral Small 2505 costs 40× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, Devstral Small 2505 or Kimi K3?
Devstral Small 2505 is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Kimi K3 lists at $3 and $15.
Is Devstral Small 2505 or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 38.9 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 128K.
How many benchmarks do Devstral Small 2505 and Kimi K3 share?
2 benchmarks have published results for both models. Devstral Small 2505 has 4 scored results on Noometry and Kimi K3 has 53.