Noometry

Model comparison

GPT-5.5 vs Qwen3 235B-A22B

GPT-5.5 is the stronger model overall, scoring 63.4 to 43.2 on the Noometry Index. Qwen3 235B-A22B costs 9.2× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.

Last verified . 38 shared benchmarks.

GPT-5.5 OpenAI

63.4

Rank #8 Confirmed

Qwen3 235B-A22B Alibaba (Qwen)

43.2

Rank #83 Confirmed

Summary

  • They share 38 benchmarks with published results for both. GPT-5.5 scores higher in 9 categories and Qwen3 235B-A22B in 0 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-5.5 leads 72.7 to 15.8.
  • The biggest single-benchmark swing is ARC-AGI-1: 95% for GPT-5.5 and 11% for Qwen3 235B-A22B.
  • Qwen3 235B-A22B is cheaper at $0.70 / $2.80 per million input/output tokens, against $5 / $30 for GPT-5.5.
  • GPT-5.5 accepts more context: 1.05M tokens versus 131K.
  • Qwen3 235B-A22B has downloadable open weights; the other is API-only.

Side by side

GPT-5.5Qwen3 235B-A22B
ProviderOpenAIAlibaba (Qwen)
Noometry Index63.443.2
Released2026-04-232025-04
WeightsProprietaryOpen
Context window1.05M131K
Max output128K16K
Input $ / M tokens$5$0.70
Output $ / M tokens$30$2.80
Results tracked7149

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

Category by category

Coding GPT-5.5 leads

GPT-5.5: 58.3 (#13), Qwen3 235B-A22B: 43.8 (#72)

BenchmarkGPT-5.5Qwen3 235B-A22B
SciCode56.1%42.4%
WeirdML84.9%41%
LMArena Coding14941445
SWE-bench Verified80.6%—
DeepSWE67%—
FrontierCode43%—
Aider Polyglot—59.6%
LMArena WebDev1512—
GSO40.2%—
MirrorCode10%—
ALE-Bench1,943—

Agentic & Tool Use GPT-5.5 leads

GPT-5.5: 50.7 (#6), Qwen3 235B-A22B: 33.7 (#48)

BenchmarkGPT-5.5Qwen3 235B-A22B
Vending-Bench 27,524-11.34
Terminal-Bench84.7%—
APEX-Agents55.1%—
Berkeley Function Calling Leaderboard—52.1%
OSWorld 2.013%—
Remote Labor Index6.3%—
τ²-bench Banking44.6%—
DeepResearch Bench54%—
PostTrainBench27.2%—
ExploitBench47.4%—
GBAEval53.2%—
GDP.pdf26%—
LMArena Search1242—

Reasoning GPT-5.5 leads

GPT-5.5: 72.7 (#8), Qwen3 235B-A22B: 15.8 (#302)

BenchmarkGPT-5.5Qwen3 235B-A22B
ARC-AGI-285%1.3%
SimpleBench69%31%
Kagi LLM Benchmark88.8%69.4%
ARC-AGI-195%11%
CritPt27.1%0%
Chess Puzzles54%12%
LMArena Hard Prompts14891433
Mystery Game Puzzles56%9%
DTBench96%80.3%
LMCA54.3%29.3%
Epoch Capabilities Index159.25143.87
ForecastBench60.659.7
NYT Connections (extended)96.2%—
EBR-Bench34.3%—
Surface Evolver Bench88.1%—
Bench to the Future 30.14—

Math GPT-5.5 leads

GPT-5.5: 81.5 (#9), Qwen3 235B-A22B: 48.5 (#51)

BenchmarkGPT-5.5Qwen3 235B-A22B
OTIS Mock AIME 2024-2025100%86.7%
LMArena Math14861431
FrontierMath (Feb 2025 set)51.7%8.5%
FrontierMath Tier 4 (v1)35.4%0%
FrontierMath (Tiers 1-3)85.3%—
FrontierMath Tier 472.5%—
MathArena Final-Answer Competitions94.3%—
ProofBench50%—
Omni-MATH—71.8%
MATH Level 5—68.9%
FrontierMath Erdős0%—

Knowledge GPT-5.5 leads

GPT-5.5: 64.4 (#14), Qwen3 235B-A22B: 49.5 (#67)

BenchmarkGPT-5.5Qwen3 235B-A22B
GPQA Diamond94%80.1%
SimpleQA Verified63%40.4%
Vectara Hallucination Rate9.3%9.3%
LMArena Expert15081460
MMLU-Pro—84.4%
Confabulations—15.6%
GPQA (HELM)—72.7%

Multimodal Not comparable

GPT-5.5: 47.1 (#11), Qwen3 235B-A22B: —

BenchmarkGPT-5.5Qwen3 235B-A22B
LMArena Vision1297—
Blueprint-Bench 236.2%—
Furniture Assembly44.2%—
LMArena Document1486—

Multilingual GPT-5.5 leads

GPT-5.5: 56.5 (#19), Qwen3 235B-A22B: 52.5 (#84)

BenchmarkGPT-5.5Qwen3 235B-A22B
LMArena Non-English14671410
LMArena Chinese15351481
LMArena French14861445
LMArena German14781432
LMArena Japanese14981400
LMArena Korean14611391
LMArena Russian14731411
LMArena Spanish14691429

Instruction Following GPT-5.5 leads

GPT-5.5: 77.7 (#18), Qwen3 235B-A22B: 72.8 (#132)

BenchmarkGPT-5.5Qwen3 235B-A22B
LMArena Instruction Following14791408
IFEval—83.5%

Long Context GPT-5.5 leads

GPT-5.5: 48.3 (#12), Qwen3 235B-A22B: 46.1 (#25)

BenchmarkGPT-5.5Qwen3 235B-A22B
LMArena Longer Query14841426
Fiction.LiveBench—75%
CL-bench Life22.2%—

Writing & Preference GPT-5.5 leads

GPT-5.5: 72.8 (#13), Qwen3 235B-A22B: 59.8 (#103)

BenchmarkGPT-5.5Qwen3 235B-A22B
LMArena Text14711419
LMArena Creative Writing14551384
EQ-Bench Creative Writing18441366
LMArena Multi-Turn14771432
Short-Story Creative Writing—83%
WildBench—86.6%
EQ-Bench 41315—

Frequently asked questions

Is GPT-5.5 better than Qwen3 235B-A22B?

GPT-5.5 is the stronger model overall, scoring 63.4 to 43.2 on the Noometry Index. Qwen3 235B-A22B costs 9.2× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.

Which is cheaper, GPT-5.5 or Qwen3 235B-A22B?

Qwen3 235B-A22B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; GPT-5.5 lists at $5 and $30.

Is GPT-5.5 or Qwen3 235B-A22B better for coding?

GPT-5.5 scores higher on coding benchmarks: 58.3 versus 43.8 in the Noometry coding category.

Which has the bigger context window?

GPT-5.5 does, with 1.05M tokens against 131K.

How many benchmarks do GPT-5.5 and Qwen3 235B-A22B share?

38 benchmarks have published results for both models. GPT-5.5 has 71 scored results on Noometry and Qwen3 235B-A22B has 49.

Related comparisons

Go deeper