Model comparison
Claude 3 Haiku vs Llama 4 Maverick
Llama 4 Maverick is the stronger model overall, scoring 30.9 to 25.9 on the Noometry Index.
Last verified . 31 shared benchmarks.
MetaSummary
- They share 31 benchmarks with published results for both. Claude 3 Haiku scores higher in 2 categories and Llama 4 Maverick in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Llama 4 Maverick leads 26.0 to 9.8.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 73% for Llama 4 Maverick.
- Llama 4 Maverick has downloadable open weights; the other is API-only.
Side by side
| Claude 3 Haiku | Llama 4 Maverick | |
|---|---|---|
| Provider | Anthropic | Meta |
| Noometry Index | 25.9 | 30.9 |
| Released | 2024-03-07 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.19 |
| Output $ / M tokens | — | $0.65 |
| Results tracked | 37 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Claude 3 Haiku: 26.4 (#325), Llama 4 Maverick: 26.6 (#324)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| WeirdML | 9.8% | 24.5% |
| BigCodeBench Instruct | 39.4% | 49.7% |
| LMArena Coding | 1199 | 1302 |
| BigCodeBench Complete | 50.1% | 61.4% |
| SWE-bench Verified (bash only) | — | 21% |
| Aider Polyglot | — | 15.6% |
| SciCode | — | 33.1% |
| CadEval | 12% | — |
| ALE-Bench | — | 172.97 |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, Llama 4 Maverick: 28.2 (#91)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.3% |
Reasoning Claude 3 Haiku leads
Claude 3 Haiku: 16.3 (#307), Llama 4 Maverick: 10.1 (#342)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 55.9% |
| LMArena Hard Prompts | 1174 | 1281 |
| DTBench | 50.1% | 61.9% |
| LMCA | 8.8% | 15.9% |
| Epoch Capabilities Index | 118.35 | 132.2 |
| ForecastBench | 53.2 | 57.5 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 27.7% |
| NYT Connections (extended) | — | 8% |
| ARC-AGI-1 | — | 4.4% |
| CritPt | — | 0% |
| EnigmaEval | — | 0.6% |
| WinoGrande | 74.2% | — |
Math Llama 4 Maverick leads
Claude 3 Haiku: 9.8 (#319), Llama 4 Maverick: 26.0 (#262)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 20.6% |
| LMArena Math | 1188 | 1299 |
| MATH Level 5 | 14.9% | 73% |
| Omni-MATH | — | 42.2% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Llama 4 Maverick leads
Claude 3 Haiku: 17.3 (#285), Llama 4 Maverick: 33.4 (#204)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 36.3% | 67% |
| Confabulations | 34.2% | 22.6% |
| LMArena Expert | 1148 | 1259 |
| Humanity's Last Exam | — | 5.7% |
| MMLU-Pro | — | 81% |
| Vectara Hallucination Rate | — | 8.2% |
| GPQA (HELM) | — | 65% |
| MMLU | 73.8% | — |
Multimodal Llama 4 Maverick leads
Claude 3 Haiku: 23.6 (#128), Llama 4 Maverick: 31.6 (#105)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | 950 | 1142 |
| GeoBench | — | 52% |
| ScienceQA | 72% | — |
| SpatialViz-Bench | — | 31.8% |
Multilingual Llama 4 Maverick leads
Claude 3 Haiku: 36.0 (#243), Llama 4 Maverick: 42.2 (#195)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1178 | 1269 |
| LMArena Chinese | 1155 | 1277 |
| LMArena French | 1195 | 1259 |
| LMArena German | 1174 | 1291 |
| LMArena Japanese | 1102 | 1207 |
| LMArena Korean | 1109 | 1203 |
| LMArena Russian | 1204 | 1286 |
| LMArena Spanish | 1166 | 1293 |
Instruction Following Llama 4 Maverick leads
Claude 3 Haiku: 61.3 (#247), Llama 4 Maverick: 71.7 (#146)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| LMArena Instruction Following | 1173 | 1267 |
| IFEval | — | 90.8% |
Long Context Claude 3 Haiku leads
Claude 3 Haiku: 36.1 (#237), Llama 4 Maverick: 31.4 (#279)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| LMArena Longer Query | 1190 | 1280 |
| Fiction.LiveBench | — | 46.2% |
Writing & Preference Llama 4 Maverick leads
Claude 3 Haiku: 29.7 (#291), Llama 4 Maverick: 38.8 (#252)
| Benchmark | Claude 3 Haiku | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1195 | 1287 |
| LMArena Creative Writing | 1157 | 1267 |
| EQ-Bench Creative Writing | 717 | 860 |
| LMArena Multi-Turn | 1190 | 1289 |
| Short-Story Creative Writing | — | 62% |
| WildBench | — | 80% |
Frequently asked questions
Is Claude 3 Haiku better than Llama 4 Maverick?
Llama 4 Maverick is the stronger model overall, scoring 30.9 to 25.9 on the Noometry Index.
Is Claude 3 Haiku or Llama 4 Maverick better for coding?
They score almost the same on coding (26.4 vs 26.6); test both on your own repository before choosing.
How many benchmarks do Claude 3 Haiku and Llama 4 Maverick share?
31 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Llama 4 Maverick has 54.