Noometry

Model comparison

GPT-5.4 mini vs Qwen3.5 27B

GPT-5.4 mini is the stronger model overall, scoring 44.8 to 41.4 on the Noometry Index. Qwen3.5 27B costs 2.0× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.

Last verified . 26 shared benchmarks.

GPT-5.4 mini OpenAI

44.8

Rank #69 Confirmed

Qwen3.5 27B Alibaba (Qwen)

41.4

Rank #127 Confirmed

Summary

  • They share 26 benchmarks with published results for both. GPT-5.4 mini scores higher in 8 categories and Qwen3.5 27B in 1 category; 6 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where GPT-5.4 mini leads 51.4 to 37.4.
  • The biggest single-benchmark swing is WeirdML: 60.3% for GPT-5.4 mini and 39.5% for Qwen3.5 27B.
  • Qwen3.5 27B is cheaper at $0.30 / $2.40 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
  • GPT-5.4 mini accepts more context: 400K tokens versus 262K.
  • Qwen3.5 27B has downloadable open weights; the other is API-only.

Side by side

GPT-5.4 miniQwen3.5 27B
ProviderOpenAIAlibaba (Qwen)
Noometry Index44.841.4
Released2026-03-172026-02-23
WeightsProprietaryOpen
Context window400K262K
Max output128K66K
Input $ / M tokens$0.75$0.30
Output $ / M tokens$4.50$2.40
Results tracked4628

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

Category by category

Coding GPT-5.4 mini leads

GPT-5.4 mini: 45.0 (#67), Qwen3.5 27B: 38.6 (#159)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena WebDev13971357
WeirdML60.3%39.5%
LMArena Coding14381427
ALE-Bench1,189349.45
FrontierCode27%—
SciCode49.9%—

Agentic & Tool Use Not comparable

GPT-5.4 mini: 29.6 (#74), Qwen3.5 27B: —

BenchmarkGPT-5.4 miniQwen3.5 27B
DeepResearch Bench36.3%—
Vending-Bench 2—201.98

Reasoning GPT-5.4 mini leads

GPT-5.4 mini: 30.3 (#77), Qwen3.5 27B: 26.6 (#116)

BenchmarkGPT-5.4 miniQwen3.5 27B
NYT Connections (extended)61.8%47.9%
Thematic Generalization61.7%45.5%
LMArena Hard Prompts14241414
DTBench80%82.4%
LMCA40.8%34%
ARC-AGI-218.9%—
Kagi LLM Benchmark37.9%—
ARC-AGI-163.7%—
CritPt10%—
Chess Puzzles24%—
Mystery Game Puzzles11%—
Epoch Capabilities Index148.83—
ForecastBench57—

Math GPT-5.4 mini leads

GPT-5.4 mini: 45.4 (#63), Qwen3.5 27B: 36.8 (#103)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Math14181428
FrontierMath (Tiers 1-3)51.2%—
FrontierMath Tier 49.8%—
MathArena Final-Answer Competitions—56.7%
OTIS Mock AIME 2024-202588.9%—
ProofBench21%—
FrontierMath (Feb 2025 set)28.3%—
FrontierMath Tier 4 (v1)2.1%—

Knowledge GPT-5.4 mini leads

GPT-5.4 mini: 51.4 (#62), Qwen3.5 27B: 37.4 (#149)

BenchmarkGPT-5.4 miniQwen3.5 27B
Vectara Hallucination Rate5.5%12.1%
LMArena Expert14341428
GPQA Diamond86.9%—
SimpleQA Verified29.4%—

Multimodal Too close to call

GPT-5.4 mini: 39.0 (#51), Qwen3.5 27B: 38.7 (#56)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Vision12451240

Multilingual GPT-5.4 mini leads

GPT-5.4 mini: 52.1 (#94), Qwen3.5 27B: 51.0 (#112)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Non-English14051390
LMArena Chinese14461478
LMArena French14401409
LMArena German14071393
LMArena Japanese13751345
LMArena Korean13691359
LMArena Russian14171389
LMArena Spanish14031408

Instruction Following Too close to call

GPT-5.4 mini: 74.3 (#99), Qwen3.5 27B: 73.7 (#114)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Instruction Following14051393

Long Context Too close to call

GPT-5.4 mini: 42.9 (#107), Qwen3.5 27B: 43.1 (#103)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Longer Query14071413

Writing & Preference GPT-5.4 mini leads

GPT-5.4 mini: 64.2 (#56), Qwen3.5 27B: 59.5 (#106)

BenchmarkGPT-5.4 miniQwen3.5 27B
LMArena Text14111409
LMArena Creative Writing13701362
LMArena Multi-Turn14291410
EQ-Bench Creative Writing1665—

Frequently asked questions

Is GPT-5.4 mini better than Qwen3.5 27B?

GPT-5.4 mini is the stronger model overall, scoring 44.8 to 41.4 on the Noometry Index. Qwen3.5 27B costs 2.0× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.

Which is cheaper, GPT-5.4 mini or Qwen3.5 27B?

Qwen3.5 27B is cheaper. It lists at $0.30 per million input tokens and $2.40 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.

Is GPT-5.4 mini or Qwen3.5 27B better for coding?

GPT-5.4 mini scores higher on coding benchmarks: 45.0 versus 38.6 in the Noometry coding category.

Which has the bigger context window?

GPT-5.4 mini does, with 400K tokens against 262K.

How many benchmarks do GPT-5.4 mini and Qwen3.5 27B share?

26 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Qwen3.5 27B has 28.

Related comparisons

Go deeper