# Claude Opus 5.5 vs Phi-4

> Claude Opus 5.5 is the stronger model overall, scoring 68.6 to 31.2 on the Noometry Index. Phi-4 costs 91× less per token, which makes it the better buy when Claude Opus 5.5's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-opus-5-5-vs-phi-4
- Last updated: 2026-10-10
- Shared benchmarks: 17

## Summary

- They share 17 benchmarks with published results for both. Claude Opus 5.5 scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Opus 5.5 leads 91.8 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 100% for Claude Opus 5.5 and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $4 / $20 for Claude Opus 5.5.
- Claude Opus 5.5 accepts more context: 1M tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| Provider | Anthropic | Microsoft |
| Noometry Index | 68.6 | 31.2 |
| Rank | 3 | 279 |
| Context | 1M | 128K |
| Input $/M | $4 | $0.07 |
| Output $/M | $20 | $0.14 |
| Weights | Proprietary | Open |

## Coding

- Claude Opus 5.5: 71.9 (#3)
- Phi-4: 34.4 (#239)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Coding | 1547 | 1231 |
| FrontierCode | 54.6% | — |
| CursorBench | 57.8% | — |
| LMArena WebDev | 1813 | — |
| FrontierSWE | 62.3% | — |
| SciCode | 66.9% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| MirrorCode | 77.4% | — |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 2,147 | — |

## Agentic & Tool Use

- Claude Opus 5.5: 45.3 (#15)
- Phi-4: 22.8 (#128)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| APEX-Agents | 73.5% | — |
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |
| GDP.pdf | 30.6% | — |
| Vending-Bench 2 | 9,235 | — |

## Reasoning

- Claude Opus 5.5: 80.2 (#3)
- Phi-4: 17.7 (#291)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1535 | 1220 |
| Epoch Capabilities Index | 167.33 | 130.42 |
| ARC-AGI-2 | 93.3% | — |
| NYT Connections (extended) | 88.5% | — |
| ARC-AGI-1 | 98.5% | — |
| CritPt | 31.7% | — |
| Chess Puzzles | — | 1% |
| EBR-Bench | 71.4% | — |
| LiveBench Reasoning | — | 47.8% |
| Mystery Game Puzzles | 71% | — |
| DTBench | 98.9% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 68.2% | — |
| LiveBench | — | 41.6% |

## Math

- Claude Opus 5.5: 91.8 (#3)
- Phi-4: 20.8 (#285)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 100% | 13.8% |
| LMArena Math | 1506 | 1246 |
| FrontierMath (Tiers 1-3) | 91.2% | — |
| FrontierMath Tier 4 | 95% | — |
| ProofBench | 100% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
| FrontierMath Erdős | 2.9% | — |

## Knowledge

- Claude Opus 5.5: 66.4 (#10)
- Phi-4: 32.6 (#209)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| GPQA Diamond | 90.6% | 56.1% |
| LMArena Expert | 1547 | 1203 |
| SimpleQA Verified | 72.2% | — |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |

## Multimodal

- Claude Opus 5.5: 57.8 (#1)
- Phi-4: —

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Vision | 1321 | — |
| Blueprint-Bench 2 | 51.2% | — |
| Furniture Assembly | 83.3% | — |

## Multilingual

- Claude Opus 5.5: 59.1 (#2)
- Phi-4: 37.2 (#237)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Non-English | 1507 | 1197 |
| LMArena Chinese | 1588 | 1212 |
| LMArena French | 1514 | 1224 |
| LMArena Russian | 1520 | 1209 |
| LMArena Spanish | 1507 | 1234 |
| LMArena German | — | 1222 |
| LMArena Japanese | — | 1158 |
| LMArena Korean | — | 1151 |

## Instruction Following

- Claude Opus 5.5: 80.0 (#3)
- Phi-4: 60.4 (#251)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1537 | 1201 |
| LiveBench Instruction Following | — | 58.4% |

## Long Context

- Claude Opus 5.5: 47.1 (#19)
- Phi-4: 36.9 (#226)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1532 | 1217 |

## Writing & Preference

- Claude Opus 5.5: 78.2 (#3)
- Phi-4: 40.5 (#244)

| Benchmark | Claude Opus 5.5 | Phi-4 |
|---|---|---|
| LMArena Text | 1515 | 1217 |
| LMArena Creative Writing | 1533 | 1182 |
| LMArena Multi-Turn | 1499 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 2050 | — |
| LiveBench Language | — | 25.6% |

## FAQ

### Is Claude Opus 5.5 better than Phi-4?

Claude Opus 5.5 is the stronger model overall, scoring 68.6 to 31.2 on the Noometry Index. Phi-4 costs 91× less per token, which makes it the better buy when Claude Opus 5.5's lead doesn't matter for your workload.

### Which is cheaper, Claude Opus 5.5 or Phi-4?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Claude Opus 5.5 lists at $4 and $20.

### Is Claude Opus 5.5 or Phi-4 better for coding?

Claude Opus 5.5 scores higher on coding benchmarks: 71.9 versus 34.4 in the Noometry coding category.

### Which has the bigger context window?

Claude Opus 5.5 does, with 1M tokens against 128K.

### How many benchmarks do Claude Opus 5.5 and Phi-4 share?

17 benchmarks have published results for both models. Claude Opus 5.5 has 44 scored results on Noometry and Phi-4 has 37.