Model comparison
Claude 3.5 Haiku vs Phi 3 Mini 4k Instruct
Claude 3.5 Haiku is the stronger model overall, scoring 28.9 to 27.4 on the Noometry Index.
Last verified . 25 shared benchmarks.
MicrosoftSummary
- They share 25 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 6 categories and Phi 3 Mini 4k Instruct in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Claude 3.5 Haiku leads 63.1 to 48.0.
- The biggest single-benchmark swing is LiveBench Coding: 51.4% for Claude 3.5 Haiku and 15.5% for Phi 3 Mini 4k Instruct.
- Phi 3 Mini 4k Instruct has downloadable open weights; the other is API-only.
Side by side
| Claude 3.5 Haiku | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | Anthropic | Microsoft |
| Noometry Index | 28.9 | 27.4 |
| Released | 2024-10-22 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 49 | 35 |
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), Phi 3 Mini 4k Instruct: 26.1 (#319)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Coding | 51.4% | 15.5% |
| LMArena Coding | 1287 | 1093 |
| Aider Polyglot | 28% | — |
| SciCode | 27.4% | — |
| WeirdML | 30.7% | — |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 59% | — |
| CadEval | 32% | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
Claude 3.5 Haiku: 27.7 (#90), Phi 3 Mini 4k Instruct: —
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| BALROG | 19.3% | — |
Reasoning Claude 3.5 Haiku leads
Claude 3.5 Haiku: 17.1 (#286), Phi 3 Mini 4k Instruct: 13.6 (#326)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Reasoning | 28.1% | 26.8% |
| LMArena Hard Prompts | 1251 | 1072 |
| LiveBench Data Analysis | 48.5% | 34.7% |
| LiveBench | 43.5% | 22.4% |
| CritPt | 0% | — |
| Chess Puzzles | — | 0% |
| DTBench | 56.7% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 127.15 | — |
| HellaSwag | — | 76.7% |
| WinoGrande | — | 70.8% |
Math Phi 3 Mini 4k Instruct leads
Claude 3.5 Haiku: 13.7 (#294), Phi 3 Mini 4k Instruct: 24.5 (#258)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Math | 35.5% | 15.7% |
| LMArena Math | 1244 | 1111 |
| OTIS Mock AIME 2024-2025 | 4.3% | — |
| Omni-MATH | 22.4% | — |
| MATH Level 5 | 46.4% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Phi 3 Mini 4k Instruct leads
Claude 3.5 Haiku: 18.7 (#276), Phi 3 Mini 4k Instruct: 27.9 (#241)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1208 | 1045 |
| MMLU | 74.3% | 68.8% |
| GPQA Diamond | 38.1% | — |
| MMLU-Pro | 60.5% | — |
| Confabulations | 36.7% | — |
| GPQA (HELM) | 36.3% | — |
| ARC (AI2) Challenge | — | 84.9% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multimodal Not comparable
Claude 3.5 Haiku: 26.3 (#113), Phi 3 Mini 4k Instruct: —
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Vision | 1092 | — |
| GeoBench | 34% | — |
Multilingual Claude 3.5 Haiku leads
Claude 3.5 Haiku: 40.2 (#213), Phi 3 Mini 4k Instruct: 26.5 (#275)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1238 | 1021 |
| LMArena Chinese | 1229 | 1021 |
| LMArena French | 1262 | 1075 |
| LMArena German | 1237 | 1044 |
| LMArena Japanese | 1175 | 935 |
| LMArena Korean | 1172 | 905 |
| LMArena Russian | 1253 | 1022 |
| LMArena Spanish | 1261 | 1084 |
Instruction Following Claude 3.5 Haiku leads
Claude 3.5 Haiku: 63.1 (#229), Phi 3 Mini 4k Instruct: 48.0 (#298)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Instruction Following | 61.9% | 39.1% |
| LMArena Instruction Following | 1241 | 1053 |
| IFEval | 79.2% | — |
Long Context Claude 3.5 Haiku leads
Claude 3.5 Haiku: 38.3 (#195), Phi 3 Mini 4k Instruct: 31.7 (#271)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1261 | 1044 |
Writing & Preference Claude 3.5 Haiku leads
Claude 3.5 Haiku: 42.9 (#228), Phi 3 Mini 4k Instruct: 27.7 (#295)
| Benchmark | Claude 3.5 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1255 | 1073 |
| LMArena Creative Writing | 1233 | 1037 |
| LMArena Multi-Turn | 1265 | 1018 |
| LiveBench Language | 35.4% | 9.2% |
| Short-Story Creative Writing | 73.5% | — |
| EQ-Bench Creative Writing | 1146 | — |
| WildBench | 76% | — |
Frequently asked questions
Is Claude 3.5 Haiku better than Phi 3 Mini 4k Instruct?
Claude 3.5 Haiku is the stronger model overall, scoring 28.9 to 27.4 on the Noometry Index.
Is Claude 3.5 Haiku or Phi 3 Mini 4k Instruct better for coding?
Claude 3.5 Haiku scores higher on coding benchmarks: 32.6 versus 26.1 in the Noometry coding category.
How many benchmarks do Claude 3.5 Haiku and Phi 3 Mini 4k Instruct share?
25 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.