# Phi-4 vs Qwen3-Next 80B-A3B Instruct

> Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 30.9 on the Noometry Index. Phi-4 costs 10.0× less per token, which makes it the better buy when Qwen3-Next 80B-A3B Instruct's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/phi-4-vs-qwen3-next-80b-a3b-instruct
- Last updated: 2026-10-10
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. Phi-4 scores higher in 0 categories and Qwen3-Next 80B-A3B Instruct in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3-Next 80B-A3B Instruct leads 58.2 to 40.7.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 3.7% for Phi-4 and 9.3% for Qwen3-Next 80B-A3B Instruct.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.50 / $2 for Qwen3-Next 80B-A3B Instruct.
- Qwen3-Next 80B-A3B Instruct accepts more context: 131K tokens versus 128K.

## Snapshot

| | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| Provider | Microsoft | Alibaba (Qwen) |
| Noometry Index | 30.9 | 42.5 |
| Rank | 269 | 100 |
| Context | 128K | 131K |
| Input $/M | $0.07 | $0.50 |
| Output $/M | $0.14 | $2 |
| Weights | Open | Open |

## Coding

- Phi-4: 33.8 (#237)
- Qwen3-Next 80B-A3B Instruct: 41.9 (#101)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Coding | 1231 | 1441 |
| BigCodeBench Instruct | 45.5% | — |
| LiveBench Coding | 30.7% | — |
| BigCodeBench Complete | 55.4% | — |

## Agentic & Tool Use

- Phi-4: 22.6 (#120)
- Qwen3-Next 80B-A3B Instruct: —

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |

## Reasoning

- Phi-4: 17.0 (#287)
- Qwen3-Next 80B-A3B Instruct: 30.1 (#79)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1428 |
| Kagi LLM Benchmark | — | 66.7% |
| Chess Puzzles | 1% | — |
| LiveBench Reasoning | 47.8% | — |
| LiveBench Data Analysis | 45.2% | — |
| Epoch Capabilities Index | 130.43 | — |
| LiveBench | 41.6% | — |

## Math

- Phi-4: 19.6 (#279)
- Qwen3-Next 80B-A3B Instruct: 36.3 (#118)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Math | 1246 | 1440 |
| OTIS Mock AIME 2024-2025 | 13.8% | — |
| Omni-MATH | — | 46.7% |
| LiveBench Math | 42% | — |
| MATH Level 5 | 64.9% | — |

## Knowledge

- Phi-4: 32.5 (#204)
- Qwen3-Next 80B-A3B Instruct: 41.3 (#102)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| Vectara Hallucination Rate | 3.7% | 9.3% |
| LMArena Expert | 1203 | 1416 |
| GPQA Diamond | 56.1% | — |
| MMLU-Pro | — | 78.6% |
| Confabulations | 29.4% | — |
| GPQA (HELM) | — | 63% |
| MMLU | 84.8% | — |

## Multilingual

- Phi-4: 37.4 (#232)
- Qwen3-Next 80B-A3B Instruct: 52.3 (#88)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Non-English | 1197 | 1407 |
| LMArena Chinese | 1211 | 1461 |
| LMArena French | 1223 | 1415 |
| LMArena German | 1222 | 1416 |
| LMArena Japanese | 1158 | 1398 |
| LMArena Korean | 1151 | 1366 |
| LMArena Russian | 1209 | 1403 |
| LMArena Spanish | 1234 | 1436 |

## Instruction Following

- Phi-4: 60.6 (#246)
- Qwen3-Next 80B-A3B Instruct: 71.0 (#154)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Instruction Following | 1201 | 1388 |
| LiveBench Instruction Following | 58.4% | — |
| IFEval | — | 81% |

## Long Context

- Phi-4: 36.9 (#221)
- Qwen3-Next 80B-A3B Instruct: 37.0 (#218)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Longer Query | 1217 | 1403 |
| Fiction.LiveBench | — | 55.6% |

## Writing & Preference

- Phi-4: 40.7 (#239)
- Qwen3-Next 80B-A3B Instruct: 58.2 (#116)

| Benchmark | Phi-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Text | 1217 | 1417 |
| LMArena Creative Writing | 1182 | 1333 |
| LMArena Multi-Turn | 1206 | 1415 |
| Short-Story Creative Writing | 62.6% | — |
| WildBench | — | 80.7% |
| LiveBench Language | 25.6% | — |

## FAQ

### Is Phi-4 better than Qwen3-Next 80B-A3B Instruct?

Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 30.9 on the Noometry Index. Phi-4 costs 10.0× less per token, which makes it the better buy when Qwen3-Next 80B-A3B Instruct's lead doesn't matter for your workload.

### Which is cheaper, Phi-4 or Qwen3-Next 80B-A3B Instruct?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Qwen3-Next 80B-A3B Instruct lists at $0.50 and $2.

### Is Phi-4 or Qwen3-Next 80B-A3B Instruct better for coding?

Qwen3-Next 80B-A3B Instruct scores higher on coding benchmarks: 41.9 versus 33.8 in the Noometry coding category.

### Which has the bigger context window?

Qwen3-Next 80B-A3B Instruct does, with 131K tokens against 128K.

### How many benchmarks do Phi-4 and Qwen3-Next 80B-A3B Instruct share?

18 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Qwen3-Next 80B-A3B Instruct has 25.