# Claude 3 Sonnet vs Phi-4

> Phi-4 is the stronger model overall, scoring 30.9 to 28.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-sonnet-vs-phi-4
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. Claude 3 Sonnet scores higher in 4 categories and Phi-4 in 4 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Phi-4 leads 32.5 to 21.1.
- The biggest single-benchmark swing is MATH Level 5: 18.2% for Claude 3 Sonnet and 64.9% for Phi-4.
- Phi-4 has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| Provider | Anthropic | Microsoft |
| Noometry Index | 28.8 | 30.9 |
| Rank | 311 | 269 |
| Context | — | 128K |
| Input $/M | — | $0.07 |
| Output $/M | — | $0.14 |
| Weights | Proprietary | Open |

## Coding

- Claude 3 Sonnet: 29.2 (#296)
- Phi-4: 33.8 (#237)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| BigCodeBench Instruct | 42.7% | 45.5% |
| LMArena Coding | 1223 | 1231 |
| BigCodeBench Complete | 53.8% | 55.4% |
| WeirdML | 10.2% | — |
| LiveBench Coding | — | 30.7% |
| HumanEval+ | 64% | — |
| MBPP+ | 69.3% | — |

## Agentic & Tool Use

- Claude 3 Sonnet: —
- Phi-4: 22.6 (#120)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |

## Reasoning

- Claude 3 Sonnet: 19.7 (#232)
- Phi-4: 17.0 (#287)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1197 | 1220 |
| Epoch Capabilities Index | 120.69 | 130.43 |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 53.6% | — |
| LiveBench Data Analysis | — | 45.2% |
| LiveBench | — | 41.6% |
| WinoGrande | 75.1% | — |

## Math

- Claude 3 Sonnet: 10.0 (#305)
- Phi-4: 19.6 (#279)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.5% | 13.8% |
| LMArena Math | 1213 | 1246 |
| MATH Level 5 | 18.2% | 64.9% |
| LiveBench Math | — | 42% |

## Knowledge

- Claude 3 Sonnet: 21.1 (#271)
- Phi-4: 32.5 (#204)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| GPQA Diamond | 40.6% | 56.1% |
| LMArena Expert | 1173 | 1203 |
| MMLU | 75.9% | 84.8% |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |

## Multimodal

- Claude 3 Sonnet: 24.7 (#120)
- Phi-4: —

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Vision | 984 | — |

## Multilingual

- Claude 3 Sonnet: 38.0 (#229)
- Phi-4: 37.4 (#232)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Non-English | 1205 | 1197 |
| LMArena Chinese | 1188 | 1211 |
| LMArena French | 1228 | 1223 |
| LMArena German | 1204 | 1222 |
| LMArena Japanese | 1131 | 1158 |
| LMArena Korean | 1128 | 1151 |
| LMArena Russian | 1227 | 1209 |
| LMArena Spanish | 1204 | 1234 |

## Instruction Following

- Claude 3 Sonnet: 63.1 (#230)
- Phi-4: 60.6 (#246)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1200 | 1201 |
| LiveBench Instruction Following | — | 58.4% |

## Long Context

- Claude 3 Sonnet: 36.7 (#223)
- Phi-4: 36.9 (#221)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1211 | 1217 |

## Writing & Preference

- Claude 3 Sonnet: 42.3 (#233)
- Phi-4: 40.7 (#239)

| Benchmark | Claude 3 Sonnet | Phi-4 |
|---|---|---|
| LMArena Text | 1218 | 1217 |
| LMArena Creative Writing | 1186 | 1182 |
| LMArena Multi-Turn | 1227 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| LiveBench Language | — | 25.6% |

## FAQ

### Is Claude 3 Sonnet better than Phi-4?

Phi-4 is the stronger model overall, scoring 30.9 to 28.8 on the Noometry Index.

### Is Claude 3 Sonnet or Phi-4 better for coding?

Phi-4 scores higher on coding benchmarks: 33.8 versus 29.2 in the Noometry coding category.

### How many benchmarks do Claude 3 Sonnet and Phi-4 share?

24 benchmarks have published results for both models. Claude 3 Sonnet has 30 scored results on Noometry and Phi-4 has 37.