Model comparison
Claude Haiku 4.5 vs Devstral Small 2505
Claude Haiku 4.5 is the stronger model overall, scoring 39.5 to 34.3 on the Noometry Index. Devstral Small 2505 costs 13× less per token, which makes it the better buy when Claude Haiku 4.5's lead doesn't matter for your workload.
Last verified . 3 shared benchmarks.
Mistral AISummary
- They share 3 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 1 category and Devstral Small 2505 in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where Claude Haiku 4.5 leads 44.0 to 38.9.
- The biggest single-benchmark swing is SciCode: 43.3% for Claude Haiku 4.5 and 28.8% for Devstral Small 2505.
- Devstral Small 2505 is cheaper at $0.10 / $0.30 per million input/output tokens, against $1 / $5 for Claude Haiku 4.5.
- Claude Haiku 4.5 accepts more context: 200K tokens versus 128K.
- Devstral Small 2505 has downloadable open weights; the other is API-only.
Side by side
| Claude Haiku 4.5 | Devstral Small 2505 | |
|---|---|---|
| Provider | Anthropic | Mistral AI |
| Noometry Index | 39.5 | 34.3 |
| Released | 2025-10-15 | 2025-05-07 |
| Weights | Proprietary | Open |
| Context window | 200K | 128K |
| Max output | 64K | 128K |
| Input $ / M tokens | $1 | $0.10 |
| Output $ / M tokens | $5 | $0.30 |
| Results tracked | 53 | 4 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Haiku 4.5 leads
Claude Haiku 4.5: 44.0 (#78), Devstral Small 2505: 38.9 (#166)
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| SWE-bench Verified (bash only) | 66.6% | 56.4% |
| SciCode | 43.3% | 28.8% |
| LMArena WebDev | 1330 | — |
| SWE-bench Multilingual | 64.7% | — |
| WeirdML | 45.4% | — |
| LMArena Coding | 1453 | — |
| ALE-Bench | 653.48 | — |
Agentic & Tool Use Not comparable
Claude Haiku 4.5: 33.6 (#52), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| Terminal-Bench | 35.5% | — |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| Vending-Bench 2 | 458.89 | — |
Reasoning Devstral Small 2505 leads
Claude Haiku 4.5: 15.1 (#320), Devstral Small 2505: 19.7 (#252)
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| CritPt | 0% | 0% |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | — | 37.7% |
| NYT Connections (extended) | 14.3% | — |
| ARC-AGI-1 | 47.7% | — |
| Chess Puzzles | 8% | — |
| LMArena Hard Prompts | 1420 | — |
| DTBench | 73.6% | — |
| LMCA | 30.9% | — |
| Epoch Capabilities Index | 142.41 | — |
| ForecastBench | 61.4 | — |
Math Not comparable
Claude Haiku 4.5: 44.9 (#78), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | — |
| Omni-MATH | 56.1% | — |
| LMArena Math | 1396 | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Not comparable
Claude Haiku 4.5: 37.7 (#153), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| GPQA Diamond | 71.2% | — |
| SimpleQA Verified | 13.2% | — |
| MMLU-Pro | 77.7% | — |
| Vectara Hallucination Rate | 9.8% | — |
| GPQA (HELM) | 60.5% | — |
| LMArena Expert | 1442 | — |
Multimodal Not comparable
Claude Haiku 4.5: 26.8 (#118), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual Not comparable
Claude Haiku 4.5: 49.9 (#129), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| LMArena Non-English | 1377 | — |
| LMArena Chinese | 1417 | — |
| LMArena French | 1408 | — |
| LMArena German | 1375 | — |
| LMArena Japanese | 1339 | — |
| LMArena Korean | 1347 | — |
| LMArena Russian | 1381 | — |
| LMArena Spanish | 1420 | — |
Instruction Following Not comparable
Claude Haiku 4.5: 71.4 (#149), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| IFEval | 80.1% | — |
| LMArena Instruction Following | 1414 | — |
Long Context Not comparable
Claude Haiku 4.5: 43.6 (#92), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| LMArena Longer Query | 1427 | — |
Writing & Preference Not comparable
Claude Haiku 4.5: 57.9 (#123), Devstral Small 2505: —
| Benchmark | Claude Haiku 4.5 | Devstral Small 2505 |
|---|---|---|
| LMArena Text | 1396 | — |
| LMArena Creative Writing | 1372 | — |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
| LMArena Multi-Turn | 1409 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than Devstral Small 2505?
Claude Haiku 4.5 is the stronger model overall, scoring 39.5 to 34.3 on the Noometry Index. Devstral Small 2505 costs 13× less per token, which makes it the better buy when Claude Haiku 4.5's lead doesn't matter for your workload.
Which is cheaper, Claude Haiku 4.5 or Devstral Small 2505?
Devstral Small 2505 is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Claude Haiku 4.5 lists at $1 and $5.
Is Claude Haiku 4.5 or Devstral Small 2505 better for coding?
Claude Haiku 4.5 scores higher on coding benchmarks: 44.0 versus 38.9 in the Noometry coding category.
Which has the bigger context window?
Claude Haiku 4.5 does, with 200K tokens against 128K.
How many benchmarks do Claude Haiku 4.5 and Devstral Small 2505 share?
3 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and Devstral Small 2505 has 4.