# Phi-4 vs Qwen3.8 27B

> Qwen3.8 27B is the stronger model overall, scoring 46.0 to 31.2 on the Noometry Index. Phi-4 costs 6.9× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/phi-4-vs-qwen3-8-27b
- Last updated: 2026-10-11
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. Phi-4 scores higher in 0 categories and Qwen3.8 27B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.8 27B leads 65.8 to 40.5.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.04 / $2.30 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 128K.

## Snapshot

| | Phi-4 | Qwen3.8 27B |
|---|---|---|
| Provider | Microsoft | Alibaba (Qwen) |
| Noometry Index | 31.2 | 46.0 |
| Rank | 279 | 68 |
| Context | 128K | 262K |
| Input $/M | $0.07 | $0.04 |
| Output $/M | $0.14 | $2.30 |
| Weights | Open | Open |

## Coding

- Phi-4: 34.4 (#239)
- Qwen3.8 27B: 50.5 (#44)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1231 | 1482 |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
| BigCodeBench Instruct | 45.5% | — |
| LiveBench Coding | 30.7% | — |
| BigCodeBench Complete | 55.4% | — |

## Agentic & Tool Use

- Phi-4: 22.8 (#128)
- Qwen3.8 27B: 32.9 (#57)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |

## Reasoning

- Phi-4: 17.7 (#291)
- Qwen3.8 27B: 41.0 (#54)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1460 |
| Epoch Capabilities Index | 130.42 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| Chess Puzzles | 1% | — |
| LiveBench Reasoning | 47.8% | — |
| DTBench | — | 88% |
| LiveBench Data Analysis | 45.2% | — |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| LiveBench | 41.6% | — |

## Math

- Phi-4: 20.8 (#285)
- Qwen3.8 27B: 37.1 (#161)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1246 | 1456 |
| OTIS Mock AIME 2024-2025 | 13.8% | — |
| ProofBench | — | 16% |
| LiveBench Math | 42% | — |
| MATH Level 5 | 64.9% | — |

## Knowledge

- Phi-4: 32.6 (#209)
- Qwen3.8 27B: 41.6 (#109)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1203 | 1482 |
| GPQA Diamond | 56.1% | — |
| Confabulations | 29.4% | — |
| Vectara Hallucination Rate | 3.7% | — |
| MMLU | 84.8% | — |

## Multimodal

- Phi-4: —
- Qwen3.8 27B: 41.3 (#37)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |

## Multilingual

- Phi-4: 37.2 (#237)
- Qwen3.8 27B: 53.7 (#60)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1197 | 1430 |
| LMArena Chinese | 1212 | 1504 |
| LMArena French | 1224 | 1465 |
| LMArena German | 1222 | 1438 |
| LMArena Japanese | 1158 | 1384 |
| LMArena Korean | 1151 | 1393 |
| LMArena Russian | 1209 | 1415 |
| LMArena Spanish | 1234 | 1448 |

## Instruction Following

- Phi-4: 60.4 (#251)
- Qwen3.8 27B: 75.8 (#53)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1201 | 1439 |
| LiveBench Instruction Following | 58.4% | — |

## Long Context

- Phi-4: 36.9 (#226)
- Qwen3.8 27B: 44.3 (#70)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1217 | 1450 |

## Writing & Preference

- Phi-4: 40.5 (#244)
- Qwen3.8 27B: 65.8 (#43)

| Benchmark | Phi-4 | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1217 | 1441 |
| LMArena Creative Writing | 1182 | 1384 |
| LMArena Multi-Turn | 1206 | 1441 |
| Short-Story Creative Writing | 62.6% | — |
| EQ-Bench Creative Writing | — | 1671 |
| LiveBench Language | 25.6% | — |

## FAQ

### Is Phi-4 better than Qwen3.8 27B?

Qwen3.8 27B is the stronger model overall, scoring 46.0 to 31.2 on the Noometry Index. Phi-4 costs 6.9× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.

### Which is cheaper, Phi-4 or Qwen3.8 27B?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Qwen3.8 27B lists at $0.04 and $2.30.

### Is Phi-4 or Qwen3.8 27B better for coding?

Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 34.4 in the Noometry coding category.

### Which has the bigger context window?

Qwen3.8 27B does, with 262K tokens against 128K.

### How many benchmarks do Phi-4 and Qwen3.8 27B share?

18 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Qwen3.8 27B has 31.