Model comparison
Claude 3 Haiku vs GPT-4o
GPT-4o is the stronger model overall, scoring 28.6 to 25.9 on the Noometry Index.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. Claude 3 Haiku scores higher in 2 categories and GPT-4o in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-4o leads 52.6 to 29.7.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 53.3% for GPT-4o.
Side by side
| Claude 3 Haiku | GPT-4o | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 25.9 | 28.6 |
| Released | 2024-03-07 | 2024-05-13 |
| Weights | Proprietary | Proprietary |
| Context window | — | 128K |
| Max output | — | 16K |
| Input $ / M tokens | — | $2.50 |
| Output $ / M tokens | — | $10 |
| Results tracked | 37 | 72 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude 3 Haiku leads
Claude 3 Haiku: 26.4 (#325), GPT-4o: 24.8 (#328)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| WeirdML | 9.8% | 25.1% |
| BigCodeBench Instruct | 39.4% | 51.1% |
| LMArena Coding | 1199 | 1297 |
| BigCodeBench Complete | 50.1% | 61.1% |
| CadEval | 12% | 26% |
| HumanEval+ | 68.9% | 87.2% |
| MBPP+ | 68.8% | 72.2% |
| SWE-bench Verified | — | 31% |
| SWE-bench Verified (bash only) | — | 21.6% |
| Aider Polyglot | — | 45.3% |
| GSO | — | 0% |
| LiveBench Coding | — | 51.4% |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, GPT-4o: 21.0 (#141)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| GDPval | — | 9.9% |
| TheAgentCompany | — | 8.6% |
| Cybench | — | 12.5% |
| BALROG | — | 32.3% |
| LMArena Search | — | 1006 |
| METR Time Horizons | — | 40.8% |
Reasoning Claude 3 Haiku leads
Claude 3 Haiku: 16.3 (#307), GPT-4o: 9.4 (#343)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1281 |
| DTBench | 50.1% | 64.5% |
| LMCA | 8.8% | 16.6% |
| Epoch Capabilities Index | 118.35 | 128.97 |
| ForecastBench | 53.2 | 57.7 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 17.8% |
| Kagi LLM Benchmark | 34.2% | — |
| ARC-AGI-1 | — | 4.5% |
| CritPt | — | 0% |
| Chess Puzzles | — | 13% |
| EnigmaEval | — | 0.8% |
| LiveBench Reasoning | — | 55.8% |
| LiveBench Data Analysis | — | 60.9% |
| LiveBench | — | 55.3% |
| WinoGrande | 74.2% | — |
Math Too close to call
Claude 3 Haiku: 9.8 (#319), GPT-4o: 10.6 (#312)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 6.4% |
| LMArena Math | 1188 | 1285 |
| MATH Level 5 | 14.9% | 53.3% |
| FrontierMath (Tiers 1-3) | — | 0.4% |
| Omni-MATH | — | 29.3% |
| LiveBench Math | — | 49.5% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge GPT-4o leads
Claude 3 Haiku: 17.3 (#285), GPT-4o: 28.8 (#242)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| GPQA Diamond | 36.3% | 49.2% |
| Confabulations | 34.2% | 15.3% |
| LMArena Expert | 1148 | 1250 |
| MMLU | 73.8% | 88.1% |
| Humanity's Last Exam | — | 2.7% |
| SimpleQA Verified | — | 26% |
| MMLU-Pro | — | 71.3% |
| Vectara Hallucination Rate | — | 9.6% |
| GPQA (HELM) | — | 52% |
Multimodal GPT-4o leads
Claude 3 Haiku: 23.6 (#128), GPT-4o: 34.5 (#91)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Vision | 950 | 1137 |
| ScienceQA | 72% | 88.5% |
| Video-MME | — | 71.9% |
| GeoBench | — | 71% |
| VPCT | — | 40% |
Multilingual GPT-4o leads
Claude 3 Haiku: 36.0 (#243), GPT-4o: 43.2 (#186)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Non-English | 1178 | 1283 |
| LMArena Chinese | 1155 | 1277 |
| LMArena French | 1195 | 1304 |
| LMArena German | 1174 | 1282 |
| LMArena Japanese | 1102 | 1257 |
| LMArena Korean | 1109 | 1234 |
| LMArena Russian | 1204 | 1286 |
| LMArena Spanish | 1166 | 1292 |
Instruction Following GPT-4o leads
Claude 3 Haiku: 61.3 (#247), GPT-4o: 66.6 (#207)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Instruction Following | 1173 | 1278 |
| LiveBench Instruction Following | — | 68.6% |
| IFEval | — | 81.7% |
Long Context GPT-4o leads
Claude 3 Haiku: 36.1 (#237), GPT-4o: 39.4 (#179)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Longer Query | 1190 | 1289 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference GPT-4o leads
Claude 3 Haiku: 29.7 (#291), GPT-4o: 52.6 (#166)
| Benchmark | Claude 3 Haiku | GPT-4o |
|---|---|---|
| LMArena Text | 1195 | 1300 |
| LMArena Creative Writing | 1157 | 1292 |
| LMArena Multi-Turn | 1190 | 1302 |
| Short-Story Creative Writing | — | 81.8% |
| EQ-Bench Creative Writing | 717 | — |
| WildBench | — | 82.8% |
| LiveBench Language | — | 47.6% |
Frequently asked questions
Is Claude 3 Haiku better than GPT-4o?
GPT-4o is the stronger model overall, scoring 28.6 to 25.9 on the Noometry Index.
Is Claude 3 Haiku or GPT-4o better for coding?
Claude 3 Haiku scores higher on coding benchmarks: 26.4 versus 24.8 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and GPT-4o share?
34 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GPT-4o has 72.