Model comparison
Claude 3 Haiku vs Claude 3 Sonnet
Claude 3 Sonnet is the stronger model overall, scoring 28.8 to 25.7 on the Noometry Index.
Last verified . 30 shared benchmarks.
Summary
- They share 30 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and Claude 3 Sonnet in 9 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Claude 3 Sonnet leads 42.3 to 29.8.
Side by side
| Claude 3 Haiku | Claude 3 Sonnet | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 25.7 | 28.8 |
| Released | 2024-03-07 | 2024-02-29 |
| Weights | Proprietary | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 37 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude 3 Sonnet leads
Claude 3 Haiku: 26.1 (#320), Claude 3 Sonnet: 29.2 (#296)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| WeirdML | 9.8% | 10.2% |
| BigCodeBench Instruct | 39.4% | 42.7% |
| LMArena Coding | 1199 | 1223 |
| BigCodeBench Complete | 50.1% | 53.8% |
| HumanEval+ | 68.9% | 64% |
| MBPP+ | 68.8% | 69.3% |
| CadEval | 12% | — |
Reasoning Claude 3 Sonnet leads
Claude 3 Haiku: 15.7 (#303), Claude 3 Sonnet: 19.7 (#232)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1197 |
| DTBench | 50.1% | 53.6% |
| Epoch Capabilities Index | 118.3 | 120.69 |
| WinoGrande | 74.2% | 75.1% |
| Kagi LLM Benchmark | 34.2% | — |
| LMCA | 8.8% | — |
| ForecastBench | 53.2 | — |
Math Too close to call
Claude 3 Haiku: 9.2 (#313), Claude 3 Sonnet: 10.0 (#305)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 2.5% |
| LMArena Math | 1188 | 1213 |
| MATH Level 5 | 14.9% | 18.2% |
Knowledge Claude 3 Sonnet leads
Claude 3 Haiku: 17.3 (#280), Claude 3 Sonnet: 21.1 (#271)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| GPQA Diamond | 36.3% | 40.6% |
| LMArena Expert | 1148 | 1173 |
| MMLU | 73.8% | 75.9% |
| Confabulations | 34.2% | — |
Multimodal Claude 3 Sonnet leads
Claude 3 Haiku: 23.1 (#123), Claude 3 Sonnet: 24.7 (#120)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Vision | 950 | 984 |
| ScienceQA | 72% | — |
Multilingual Claude 3 Sonnet leads
Claude 3 Haiku: 36.1 (#238), Claude 3 Sonnet: 38.0 (#229)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Non-English | 1178 | 1205 |
| LMArena Chinese | 1155 | 1188 |
| LMArena French | 1194 | 1228 |
| LMArena German | 1174 | 1204 |
| LMArena Japanese | 1102 | 1131 |
| LMArena Korean | 1108 | 1128 |
| LMArena Russian | 1204 | 1227 |
| LMArena Spanish | 1165 | 1204 |
Instruction Following Claude 3 Sonnet leads
Claude 3 Haiku: 61.5 (#242), Claude 3 Sonnet: 63.1 (#230)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Instruction Following | 1174 | 1200 |
Long Context Too close to call
Claude 3 Haiku: 36.1 (#232), Claude 3 Sonnet: 36.7 (#223)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Longer Query | 1191 | 1211 |
Writing & Preference Claude 3 Sonnet leads
Claude 3 Haiku: 29.8 (#286), Claude 3 Sonnet: 42.3 (#233)
| Benchmark | Claude 3 Haiku | Claude 3 Sonnet |
|---|---|---|
| LMArena Text | 1195 | 1218 |
| LMArena Creative Writing | 1157 | 1186 |
| LMArena Multi-Turn | 1190 | 1227 |
| EQ-Bench Creative Writing | 717 | — |
Frequently asked questions
Is Claude 3 Haiku better than Claude 3 Sonnet?
Claude 3 Sonnet is the stronger model overall, scoring 28.8 to 25.7 on the Noometry Index.
Is Claude 3 Haiku or Claude 3 Sonnet better for coding?
Claude 3 Sonnet scores higher on coding benchmarks: 29.2 versus 26.1 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and Claude 3 Sonnet share?
30 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Claude 3 Sonnet has 30.