Model comparison
Claude 3.5 Haiku vs Claude 3 Haiku
Claude 3.5 Haiku is the stronger model overall, scoring 29.2 to 25.9 on the Noometry Index.
Last verified . 30 shared benchmarks.
Summary
- They share 30 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 9 categories and Claude 3 Haiku in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Claude 3.5 Haiku leads 42.7 to 29.7.
- The biggest single-benchmark swing is MATH Level 5: 46.4% for Claude 3.5 Haiku and 14.9% for Claude 3 Haiku.
Side by side
| Claude 3.5 Haiku | Claude 3 Haiku | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 29.2 | 25.9 |
| Released | 2024-10-22 | 2024-03-07 |
| Weights | Proprietary | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 49 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude 3.5 Haiku leads
Claude 3.5 Haiku: 32.9 (#265), Claude 3 Haiku: 26.4 (#325)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| WeirdML | 30.7% | 9.8% |
| BigCodeBench Instruct | 46.1% | 39.4% |
| LMArena Coding | 1286 | 1199 |
| BigCodeBench Complete | 59% | 50.1% |
| CadEval | 32% | 12% |
| Aider Polyglot | 28% | — |
| SciCode | 27.4% | — |
| LiveBench Coding | 51.4% | — |
| HumanEval+ | — | 68.9% |
| MBPP+ | — | 68.8% |
Agentic & Tool Use Not comparable
Claude 3.5 Haiku: 28.0 (#95), Claude 3 Haiku: —
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| BALROG | 19.3% | — |
Reasoning Claude 3.5 Haiku leads
Claude 3.5 Haiku: 17.7 (#290), Claude 3 Haiku: 16.3 (#307)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1174 |
| DTBench | 56.7% | 50.1% |
| Epoch Capabilities Index | 127.15 | 118.35 |
| Kagi LLM Benchmark | — | 34.2% |
| CritPt | 0% | — |
| LiveBench Reasoning | 28.1% | — |
| LiveBench Data Analysis | 48.5% | — |
| LMCA | — | 8.8% |
| ForecastBench | — | 53.2 |
| LiveBench | 43.5% | — |
| WinoGrande | — | 74.2% |
Math Claude 3.5 Haiku leads
Claude 3.5 Haiku: 14.7 (#300), Claude 3 Haiku: 9.8 (#319)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 1.8% |
| LMArena Math | 1244 | 1188 |
| MATH Level 5 | 46.4% | 14.9% |
| Omni-MATH | 22.4% | — |
| LiveBench Math | 35.5% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Claude 3.5 Haiku leads
Claude 3.5 Haiku: 18.7 (#281), Claude 3 Haiku: 17.3 (#285)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| GPQA Diamond | 38.1% | 36.3% |
| Confabulations | 36.7% | 34.2% |
| LMArena Expert | 1208 | 1148 |
| MMLU | 74.3% | 73.8% |
| MMLU-Pro | 60.5% | — |
| GPQA (HELM) | 36.3% | — |
Multimodal Claude 3.5 Haiku leads
Claude 3.5 Haiku: 26.8 (#117), Claude 3 Haiku: 23.6 (#128)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Vision | 1092 | 950 |
| GeoBench | 34% | — |
| ScienceQA | — | 72% |
Multilingual Claude 3.5 Haiku leads
Claude 3.5 Haiku: 40.0 (#218), Claude 3 Haiku: 36.0 (#243)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Non-English | 1238 | 1178 |
| LMArena Chinese | 1229 | 1155 |
| LMArena French | 1264 | 1195 |
| LMArena German | 1237 | 1174 |
| LMArena Japanese | 1175 | 1102 |
| LMArena Korean | 1173 | 1109 |
| LMArena Russian | 1253 | 1204 |
| LMArena Spanish | 1261 | 1166 |
Instruction Following Claude 3.5 Haiku leads
Claude 3.5 Haiku: 62.9 (#234), Claude 3 Haiku: 61.3 (#247)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Instruction Following | 1241 | 1173 |
| LiveBench Instruction Following | 61.9% | — |
| IFEval | 79.2% | — |
Long Context Claude 3.5 Haiku leads
Claude 3.5 Haiku: 38.3 (#200), Claude 3 Haiku: 36.1 (#237)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Longer Query | 1261 | 1190 |
Writing & Preference Claude 3.5 Haiku leads
Claude 3.5 Haiku: 42.7 (#234), Claude 3 Haiku: 29.7 (#291)
| Benchmark | Claude 3.5 Haiku | Claude 3 Haiku |
|---|---|---|
| LMArena Text | 1255 | 1195 |
| LMArena Creative Writing | 1233 | 1157 |
| EQ-Bench Creative Writing | 1146 | 717 |
| LMArena Multi-Turn | 1265 | 1190 |
| Short-Story Creative Writing | 73.5% | — |
| WildBench | 76% | — |
| LiveBench Language | 35.4% | — |
Frequently asked questions
Is Claude 3.5 Haiku better than Claude 3 Haiku?
Claude 3.5 Haiku is the stronger model overall, scoring 29.2 to 25.9 on the Noometry Index.
Is Claude 3.5 Haiku or Claude 3 Haiku better for coding?
Claude 3.5 Haiku scores higher on coding benchmarks: 32.9 versus 26.4 in the Noometry coding category.
How many benchmarks do Claude 3.5 Haiku and Claude 3 Haiku share?
30 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and Claude 3 Haiku has 37.