Noometry

Model comparison

Phi-4 vs Qwen3.5 27B

Qwen3.5 27B is the stronger model overall, scoring 41.9 to 31.2 on the Noometry Index. Phi-4 costs 9.4× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.

Last verified . 18 shared benchmarks.

Phi-4 Microsoft

31.2

Rank #279 Confirmed

Qwen3.5 27B Alibaba (Qwen)

41.9

Rank #127 Confirmed

Summary

  • They share 18 benchmarks with published results for both. Phi-4 scores higher in 0 categories and Qwen3.5 27B in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Qwen3.5 27B leads 59.3 to 40.5.
  • The biggest single-benchmark swing is Vectara Hallucination Rate: 3.7% for Phi-4 and 12.1% for Qwen3.5 27B.
  • Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.30 / $2.40 for Qwen3.5 27B.
  • Qwen3.5 27B accepts more context: 262K tokens versus 128K.

Side by side

Phi-4Qwen3.5 27B
ProviderMicrosoftAlibaba (Qwen)
Noometry Index31.241.9
Released2024-12-112026-02-23
WeightsOpenOpen
Context window128K262K
Max output4K66K
Input $ / M tokens$0.07$0.30
Output $ / M tokens$0.14$2.40
Results tracked3728

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Qwen3.5 27B leads

Phi-4: 34.4 (#239), Qwen3.5 27B: 38.9 (#168)

BenchmarkPhi-4Qwen3.5 27B
LMArena Coding12311427
LMArena WebDev—1358
WeirdML—39.5%
BigCodeBench Instruct45.5%—
LiveBench Coding30.7%—
BigCodeBench Complete55.4%—
ALE-Bench—349.45

Agentic & Tool Use Not comparable

Phi-4: 22.8 (#128), Qwen3.5 27B: —

BenchmarkPhi-4Qwen3.5 27B
Berkeley Function Calling Leaderboard28.8%—
BALROG11.6%—
Vending-Bench 2—201.98

Reasoning Qwen3.5 27B leads

Phi-4: 17.7 (#291), Qwen3.5 27B: 27.5 (#117)

BenchmarkPhi-4Qwen3.5 27B
LMArena Hard Prompts12201414
NYT Connections (extended)—47.9%
Chess Puzzles1%—
Thematic Generalization—45.5%
LiveBench Reasoning47.8%—
DTBench—82.4%
LiveBench Data Analysis45.2%—
LMCA—34%
Epoch Capabilities Index130.42—
LiveBench41.6%—

Math Qwen3.5 27B leads

Phi-4: 20.8 (#285), Qwen3.5 27B: 38.8 (#127)

BenchmarkPhi-4Qwen3.5 27B
LMArena Math12461429
MathArena Final-Answer Competitions—56.7%
OTIS Mock AIME 2024-202513.8%—
LiveBench Math42%—
MATH Level 564.9%—

Knowledge Qwen3.5 27B leads

Phi-4: 32.6 (#209), Qwen3.5 27B: 38.0 (#150)

BenchmarkPhi-4Qwen3.5 27B
Vectara Hallucination Rate3.7%12.1%
LMArena Expert12031428
GPQA Diamond56.1%—
Confabulations29.4%—
MMLU84.8%—

Multimodal Not comparable

Phi-4: —, Qwen3.5 27B: 39.4 (#59)

BenchmarkPhi-4Qwen3.5 27B
LMArena Vision—1241

Multilingual Qwen3.5 27B leads

Phi-4: 37.2 (#237), Qwen3.5 27B: 50.8 (#115)

BenchmarkPhi-4Qwen3.5 27B
LMArena Non-English11971390
LMArena Chinese12121478
LMArena French12241410
LMArena German12221393
LMArena Japanese11581345
LMArena Korean11511358
LMArena Russian12091390
LMArena Spanish12341407

Instruction Following Qwen3.5 27B leads

Phi-4: 60.4 (#251), Qwen3.5 27B: 73.5 (#119)

BenchmarkPhi-4Qwen3.5 27B
LMArena Instruction Following12011393
LiveBench Instruction Following58.4%—

Long Context Qwen3.5 27B leads

Phi-4: 36.9 (#226), Qwen3.5 27B: 43.1 (#106)

BenchmarkPhi-4Qwen3.5 27B
LMArena Longer Query12171413

Writing & Preference Qwen3.5 27B leads

Phi-4: 40.5 (#244), Qwen3.5 27B: 59.3 (#111)

BenchmarkPhi-4Qwen3.5 27B
LMArena Text12171409
LMArena Creative Writing11821362
LMArena Multi-Turn12061410
Short-Story Creative Writing62.6%—
LiveBench Language25.6%—

Frequently asked questions

Is Phi-4 better than Qwen3.5 27B?

Qwen3.5 27B is the stronger model overall, scoring 41.9 to 31.2 on the Noometry Index. Phi-4 costs 9.4× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.

Which is cheaper, Phi-4 or Qwen3.5 27B?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Qwen3.5 27B lists at $0.30 and $2.40.

Is Phi-4 or Qwen3.5 27B better for coding?

Qwen3.5 27B scores higher on coding benchmarks: 38.9 versus 34.4 in the Noometry coding category.

Which has the bigger context window?

Qwen3.5 27B does, with 262K tokens against 128K.

How many benchmarks do Phi-4 and Qwen3.5 27B share?

18 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Qwen3.5 27B has 28.

Related comparisons

Go deeper