Noometry

Model comparison

GPT-4 vs Qwen3.5 397B-A17B

Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 29.1 on the Noometry Index.

Last verified . 25 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

Qwen3.5 397B-A17B Alibaba (Qwen)

46.0

Rank #67 Confirmed

Summary

  • They share 25 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Qwen3.5 397B-A17B in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Qwen3.5 397B-A17B leads 46.1 to 10.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 88.9% for Qwen3.5 397B-A17B.
  • Qwen3.5 397B-A17B is cheaper at $0.60 / $3.60 per million input/output tokens, against $30 / $60 for GPT-4.
  • Qwen3.5 397B-A17B accepts more context: 262K tokens versus 8K.
  • Qwen3.5 397B-A17B has downloadable open weights; the other is API-only.

Side by side

GPT-4Qwen3.5 397B-A17B
ProviderOpenAIAlibaba (Qwen)
Noometry Index29.146.0
Released2023-03-142026-02-01
WeightsProprietaryOpen
Context window8K262K
Max output8K66K
Input $ / M tokens$30$0.60
Output $ / M tokens$60$3.60
Results tracked3836

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

Category by category

Coding Qwen3.5 397B-A17B leads

GPT-4: 31.6 (#283), Qwen3.5 397B-A17B: 42.0 (#114)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Coding12541465
LMArena WebDev—1400
WeirdML12.4%—
BigCodeBench Instruct46%—
BigCodeBench Complete57.2%—
HumanEval+79.3%—

Agentic & Tool Use Not comparable

GPT-4: —, Qwen3.5 397B-A17B: 33.3 (#53)

BenchmarkGPT-4Qwen3.5 397B-A17B
APEX-Agents—24.9%
τ²-bench Airline—81.5%
τ²-bench Banking—9.8%
τ²-bench Retail—84.4%
τ²-bench Telecom—97.8%
METR Time Horizons36.1%—

Reasoning Qwen3.5 397B-A17B leads

GPT-4: 17.8 (#289), Qwen3.5 397B-A17B: 34.5 (#70)

BenchmarkGPT-4Qwen3.5 397B-A17B
Chess Puzzles4%13%
LMArena Hard Prompts12411448
Mystery Game Puzzles12%18%
DTBench62.7%87.5%
LMCA17.1%37.9%
Epoch Capabilities Index125.89146.65
Kagi LLM Benchmark—73.7%
NYT Connections (extended)—58.9%
Thematic Generalization—65.1%
BIG-Bench Hard75.1%—
ForecastBench57.8—
HellaSwag95.3%—
WinoGrande87.5%—

Math Qwen3.5 397B-A17B leads

GPT-4: 10.8 (#309), Qwen3.5 397B-A17B: 46.1 (#73)

BenchmarkGPT-4Qwen3.5 397B-A17B
OTIS Mock AIME 2024-20251.1%88.9%
LMArena Math12691454
FrontierMath (Tiers 1-3)—31.2%
MATH Level 523%—
GSM8K92%—

Knowledge Qwen3.5 397B-A17B leads

GPT-4: 18.4 (#282), Qwen3.5 397B-A17B: 53.3 (#58)

BenchmarkGPT-4Qwen3.5 397B-A17B
GPQA Diamond35.7%86.4%
LMArena Expert12111462
MMLU86.4%—
TriviaQA84.8%—

Multimodal Not comparable

GPT-4: —, Qwen3.5 397B-A17B: 40.7 (#44)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Vision—1263

Multilingual Qwen3.5 397B-A17B leads

GPT-4: 40.6 (#215), Qwen3.5 397B-A17B: 53.7 (#59)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Non-English12461430
LMArena Chinese12421500
LMArena French12831461
LMArena German12511447
LMArena Japanese12091426
LMArena Korean11841384
LMArena Russian12511429
LMArena Spanish12611441

Instruction Following Qwen3.5 397B-A17B leads

GPT-4: 65.3 (#222), Qwen3.5 397B-A17B: 75.0 (#77)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Instruction Following12411424

Long Context Qwen3.5 397B-A17B leads

GPT-4: 37.7 (#212), Qwen3.5 397B-A17B: 44.1 (#74)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Longer Query12441442

Writing & Preference Qwen3.5 397B-A17B leads

GPT-4: 34.9 (#268), Qwen3.5 397B-A17B: 62.3 (#79)

BenchmarkGPT-4Qwen3.5 397B-A17B
LMArena Text12631438
LMArena Creative Writing12441401
EQ-Bench Creative Writing7521478
LMArena Multi-Turn12571446

Frequently asked questions

Is GPT-4 better than Qwen3.5 397B-A17B?

Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 29.1 on the Noometry Index.

Which is cheaper, GPT-4 or Qwen3.5 397B-A17B?

Qwen3.5 397B-A17B is cheaper. It lists at $0.60 per million input tokens and $3.60 per million output tokens; GPT-4 lists at $30 and $60.

Is GPT-4 or Qwen3.5 397B-A17B better for coding?

Qwen3.5 397B-A17B scores higher on coding benchmarks: 42.0 versus 31.6 in the Noometry coding category.

Which has the bigger context window?

Qwen3.5 397B-A17B does, with 262K tokens against 8K.

How many benchmarks do GPT-4 and Qwen3.5 397B-A17B share?

25 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Qwen3.5 397B-A17B has 36.

Related comparisons

Go deeper