Model comparison
Claude 3 Haiku vs Phi 3 Mini 4k Instruct
Phi 3 Mini 4k Instruct is the stronger model overall, scoring 27.4 to 25.7 on the Noometry Index.
Last verified . 21 shared benchmarks.
MicrosoftSummary
- They share 21 benchmarks with published results for both. Claude 3 Haiku scores higher in 5 categories and Phi 3 Mini 4k Instruct in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Phi 3 Mini 4k Instruct leads 24.5 to 9.2.
- Phi 3 Mini 4k Instruct has downloadable open weights; the other is API-only.
Side by side
| Claude 3 Haiku | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | Anthropic | Microsoft |
| Noometry Index | 25.7 | 27.4 |
| Released | 2024-03-07 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 37 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Claude 3 Haiku: 26.1 (#320), Phi 3 Mini 4k Instruct: 26.1 (#319)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1199 | 1093 |
| HumanEval+ | 68.9% | 59.1% |
| MBPP+ | 68.8% | 54.2% |
| WeirdML | 9.8% | — |
| BigCodeBench Instruct | 39.4% | — |
| LiveBench Coding | — | 15.5% |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
Reasoning Claude 3 Haiku leads
Claude 3 Haiku: 15.7 (#303), Phi 3 Mini 4k Instruct: 13.6 (#326)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1072 |
| WinoGrande | 74.2% | 70.8% |
| Kagi LLM Benchmark | 34.2% | — |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | — | 26.8% |
| DTBench | 50.1% | — |
| LiveBench Data Analysis | — | 34.7% |
| LMCA | 8.8% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 118.3 | — |
| ForecastBench | 53.2 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
Math Phi 3 Mini 4k Instruct leads
Claude 3 Haiku: 9.2 (#313), Phi 3 Mini 4k Instruct: 24.5 (#258)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1188 | 1111 |
| OTIS Mock AIME 2024-2025 | 1.8% | — |
| LiveBench Math | — | 15.7% |
| MATH Level 5 | 14.9% | — |
Knowledge Phi 3 Mini 4k Instruct leads
Claude 3 Haiku: 17.3 (#280), Phi 3 Mini 4k Instruct: 27.9 (#241)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1148 | 1045 |
| MMLU | 73.8% | 68.8% |
| GPQA Diamond | 36.3% | — |
| Confabulations | 34.2% | — |
| ARC (AI2) Challenge | — | 84.9% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multimodal Not comparable
Claude 3 Haiku: 23.1 (#123), Phi 3 Mini 4k Instruct: —
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |
Multilingual Claude 3 Haiku leads
Claude 3 Haiku: 36.1 (#238), Phi 3 Mini 4k Instruct: 26.5 (#275)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1178 | 1021 |
| LMArena Chinese | 1155 | 1021 |
| LMArena French | 1194 | 1075 |
| LMArena German | 1174 | 1044 |
| LMArena Japanese | 1102 | 935 |
| LMArena Korean | 1108 | 905 |
| LMArena Russian | 1204 | 1022 |
| LMArena Spanish | 1165 | 1084 |
Instruction Following Claude 3 Haiku leads
Claude 3 Haiku: 61.5 (#242), Phi 3 Mini 4k Instruct: 48.0 (#298)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1174 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context Claude 3 Haiku leads
Claude 3 Haiku: 36.1 (#232), Phi 3 Mini 4k Instruct: 31.7 (#271)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1191 | 1044 |
Writing & Preference Claude 3 Haiku leads
Claude 3 Haiku: 29.8 (#286), Phi 3 Mini 4k Instruct: 27.7 (#295)
| Benchmark | Claude 3 Haiku | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1195 | 1073 |
| LMArena Creative Writing | 1157 | 1037 |
| LMArena Multi-Turn | 1190 | 1018 |
| EQ-Bench Creative Writing | 717 | — |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is Claude 3 Haiku better than Phi 3 Mini 4k Instruct?
Phi 3 Mini 4k Instruct is the stronger model overall, scoring 27.4 to 25.7 on the Noometry Index.
Is Claude 3 Haiku or Phi 3 Mini 4k Instruct better for coding?
They score almost the same on coding (26.1 vs 26.1); test both on your own repository before choosing.
How many benchmarks do Claude 3 Haiku and Phi 3 Mini 4k Instruct share?
21 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.