Model comparison
Claude 3.5 Haiku vs GPT-4o
Claude 3.5 Haiku and GPT-4o score almost the same on the Noometry Index (28.9 vs 28.5), so choose on price, context window or the category you care about most.
Last verified . 47 shared benchmarks.
Summary
- They share 47 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 4 categories and GPT-4o in 6 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-4o leads 28.8 to 18.7.
- The biggest single-benchmark swing is GeoBench: 34% for Claude 3.5 Haiku and 71% for GPT-4o.
Side by side
| Claude 3.5 Haiku | GPT-4o | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 28.9 | 28.5 |
| Released | 2024-10-22 | 2024-05-13 |
| Weights | Proprietary | Proprietary |
| Context window | — | 128K |
| Max output | — | 16K |
| Input $ / M tokens | — | $2.50 |
| Output $ / M tokens | — | $10 |
| Results tracked | 49 | 72 |
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: 32.6 (#258), GPT-4o: 25.1 (#323)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| Aider Polyglot | 28% | 45.3% |
| WeirdML | 30.7% | 25.1% |
| BigCodeBench Instruct | 46.1% | 51.1% |
| LiveBench Coding | 51.4% | 51.4% |
| LMArena Coding | 1287 | 1297 |
| BigCodeBench Complete | 59% | 61.1% |
| CadEval | 32% | 26% |
| SWE-bench Verified | — | 31% |
| SWE-bench Verified (bash only) | — | 21.6% |
| SciCode | 27.4% | — |
| GSO | — | 0% |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 72.2% |
Agentic & Tool Use Claude 3.5 Haiku leads
Claude 3.5 Haiku: 27.7 (#90), GPT-4o: 20.8 (#133)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| BALROG | 19.3% | 32.3% |
| GDPval | — | 9.9% |
| TheAgentCompany | — | 8.6% |
| Cybench | — | 12.5% |
| LMArena Search | — | 1006 |
| METR Time Horizons | — | 40.8% |
Reasoning Claude 3.5 Haiku leads
Claude 3.5 Haiku: 17.1 (#286), GPT-4o: 9.4 (#338)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| CritPt | 0% | 0% |
| LiveBench Reasoning | 28.1% | 55.8% |
| LMArena Hard Prompts | 1251 | 1281 |
| DTBench | 56.7% | 64.5% |
| LiveBench Data Analysis | 48.5% | 60.9% |
| Epoch Capabilities Index | 127.15 | 128.97 |
| LiveBench | 43.5% | 55.3% |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 17.8% |
| ARC-AGI-1 | — | 4.5% |
| Chess Puzzles | — | 13% |
| EnigmaEval | — | 0.8% |
| LMCA | — | 16.6% |
| ForecastBench | — | 57.7 |
Math Claude 3.5 Haiku leads
Claude 3.5 Haiku: 13.7 (#294), GPT-4o: 10.2 (#303)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 6.4% |
| Omni-MATH | 22.4% | 29.3% |
| LiveBench Math | 35.5% | 49.5% |
| LMArena Math | 1244 | 1285 |
| MATH Level 5 | 46.4% | 53.3% |
| FrontierMath (Feb 2025 set) | 0.3% | 0.3% |
| FrontierMath (Tiers 1-3) | — | 0.4% |
Knowledge GPT-4o leads
Claude 3.5 Haiku: 18.7 (#276), GPT-4o: 28.8 (#235)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| GPQA Diamond | 38.1% | 49.2% |
| MMLU-Pro | 60.5% | 71.3% |
| Confabulations | 36.7% | 15.3% |
| GPQA (HELM) | 36.3% | 52% |
| LMArena Expert | 1208 | 1250 |
| MMLU | 74.3% | 88.1% |
| Humanity's Last Exam | — | 2.7% |
| SimpleQA Verified | — | 26% |
| Vectara Hallucination Rate | — | 9.6% |
Multimodal GPT-4o leads
Claude 3.5 Haiku: 26.3 (#113), GPT-4o: 33.9 (#85)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| LMArena Vision | 1092 | 1137 |
| GeoBench | 34% | 71% |
| Video-MME | — | 71.9% |
| VPCT | — | 40% |
| ScienceQA | — | 88.5% |
Multilingual GPT-4o leads
Claude 3.5 Haiku: 40.2 (#213), GPT-4o: 43.4 (#181)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| LMArena Non-English | 1238 | 1283 |
| LMArena Chinese | 1229 | 1277 |
| LMArena French | 1262 | 1303 |
| LMArena German | 1237 | 1282 |
| LMArena Japanese | 1175 | 1257 |
| LMArena Korean | 1172 | 1234 |
| LMArena Russian | 1253 | 1286 |
| LMArena Spanish | 1261 | 1292 |
Instruction Following GPT-4o leads
Claude 3.5 Haiku: 63.1 (#229), GPT-4o: 66.8 (#202)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| LiveBench Instruction Following | 61.9% | 68.6% |
| IFEval | 79.2% | 81.7% |
| LMArena Instruction Following | 1241 | 1278 |
Long Context GPT-4o leads
Claude 3.5 Haiku: 38.3 (#195), GPT-4o: 39.4 (#174)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| LMArena Longer Query | 1261 | 1289 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference GPT-4o leads
Claude 3.5 Haiku: 42.9 (#228), GPT-4o: 52.8 (#160)
| Benchmark | Claude 3.5 Haiku | GPT-4o |
|---|---|---|
| LMArena Text | 1255 | 1300 |
| LMArena Creative Writing | 1233 | 1291 |
| Short-Story Creative Writing | 73.5% | 81.8% |
| WildBench | 76% | 82.8% |
| LMArena Multi-Turn | 1265 | 1302 |
| LiveBench Language | 35.4% | 47.6% |
| EQ-Bench Creative Writing | 1146 | — |
Frequently asked questions
Is Claude 3.5 Haiku better than GPT-4o?
Claude 3.5 Haiku and GPT-4o score almost the same on the Noometry Index (28.9 vs 28.5), so choose on price, context window or the category you care about most.
Is Claude 3.5 Haiku or GPT-4o better for coding?
Claude 3.5 Haiku scores higher on coding benchmarks: 32.6 versus 25.1 in the Noometry coding category.
How many benchmarks do Claude 3.5 Haiku and GPT-4o share?
47 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and GPT-4o has 72.