Model comparison
Claude 3 Haiku vs Claude Sonnet 4
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 25.7 on the Noometry Index.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. Claude 3 Haiku scores higher in 1 category and Claude Sonnet 4 in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 4 leads 41.5 to 9.2.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 84.4% for Claude Sonnet 4.
Side by side
| Claude 3 Haiku | Claude Sonnet 4 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 25.7 | 40.6 |
| Released | 2024-03-07 | 2025-05-22 |
| Weights | Proprietary | Proprietary |
| Context window | — | 200K |
| Max output | — | 64K |
| Input $ / M tokens | — | $3 |
| Output $ / M tokens | — | $15 |
| Results tracked | 37 | 58 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude 3 Haiku: 26.1 (#320), Claude Sonnet 4: 44.0 (#70)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| WeirdML | 9.8% | 46.1% |
| LMArena Coding | 1199 | 1414 |
| SWE-bench Verified (bash only) | — | 64.9% |
| Aider Polyglot | — | 61.3% |
| SciCode | — | 40% |
| GSO | — | 4.9% |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| ALE-Bench | — | 655.35 |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, Claude Sonnet 4: 38.3 (#29)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| TheAgentCompany | — | 33.1% |
| Cybench | — | 35% |
| DeepResearch Bench | — | 46.6% |
| OSWorld | — | 43.9% |
| METR Time Horizons | — | 62% |
Reasoning Claude Sonnet 4 leads
Claude 3 Haiku: 15.7 (#303), Claude Sonnet 4: 23.0 (#169)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 73% |
| LMArena Hard Prompts | 1174 | 1372 |
| DTBench | 50.1% | 77.1% |
| LMCA | 8.8% | 29% |
| Epoch Capabilities Index | 118.3 | 141.69 |
| ForecastBench | 53.2 | 60.2 |
| ARC-AGI-2 | — | 5.9% |
| SimpleBench | — | 45.5% |
| ARC-AGI-1 | — | 40% |
| CritPt | — | 0.3% |
| EnigmaEval | — | 3.1% |
| WinoGrande | 74.2% | — |
Math Claude Sonnet 4 leads
Claude 3 Haiku: 9.2 (#313), Claude Sonnet 4: 41.5 (#71)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 71.1% |
| LMArena Math | 1188 | 1375 |
| MATH Level 5 | 14.9% | 84.4% |
| Omni-MATH | — | 60.2% |
| FrontierMath (Feb 2025 set) | — | 4.1% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Claude Sonnet 4 leads
Claude 3 Haiku: 17.3 (#280), Claude Sonnet 4: 41.6 (#99)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| GPQA Diamond | 36.3% | 79.2% |
| Confabulations | 34.2% | 13.2% |
| LMArena Expert | 1148 | 1371 |
| Humanity's Last Exam | — | 7.8% |
| MMLU-Pro | — | 84.3% |
| Vectara Hallucination Rate | — | 10.3% |
| GPQA (HELM) | — | 70.6% |
| MMLU | 73.8% | — |
Multimodal Claude Sonnet 4 leads
Claude 3 Haiku: 23.1 (#123), Claude Sonnet 4: 25.6 (#116)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| LMArena Vision | 950 | 1191 |
| GeoBench | — | 37% |
| VPCT | — | 34% |
| MindCube | — | 44.8% |
| ScienceQA | 72% | — |
Multilingual Claude Sonnet 4 leads
Claude 3 Haiku: 36.1 (#238), Claude Sonnet 4: 46.9 (#151)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| LMArena Non-English | 1178 | 1333 |
| LMArena Chinese | 1155 | 1351 |
| LMArena French | 1194 | 1364 |
| LMArena German | 1174 | 1331 |
| LMArena Japanese | 1102 | 1301 |
| LMArena Korean | 1108 | 1293 |
| LMArena Russian | 1204 | 1355 |
| LMArena Spanish | 1165 | 1356 |
Instruction Following Claude Sonnet 4 leads
Claude 3 Haiku: 61.5 (#242), Claude Sonnet 4: 71.8 (#141)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| LMArena Instruction Following | 1174 | 1375 |
| IFEval | — | 84% |
Long Context Claude 3 Haiku leads
Claude 3 Haiku: 36.1 (#232), Claude Sonnet 4: 33.6 (#254)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| LMArena Longer Query | 1191 | 1398 |
| Fiction.LiveBench | — | 46.9% |
Writing & Preference Claude Sonnet 4 leads
Claude 3 Haiku: 29.8 (#286), Claude Sonnet 4: 57.3 (#127)
| Benchmark | Claude 3 Haiku | Claude Sonnet 4 |
|---|---|---|
| LMArena Text | 1195 | 1351 |
| LMArena Creative Writing | 1157 | 1345 |
| EQ-Bench Creative Writing | 717 | 1483 |
| LMArena Multi-Turn | 1190 | 1376 |
| Short-Story Creative Writing | — | 81.4% |
| WildBench | — | 83.8% |
Frequently asked questions
Is Claude 3 Haiku better than Claude Sonnet 4?
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 25.7 on the Noometry Index.
Is Claude 3 Haiku or Claude Sonnet 4 better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 44.0 versus 26.1 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and Claude Sonnet 4 share?
29 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Claude Sonnet 4 has 58.