Which is better, Claude Sonnet 4 or Qwen3.5 4B?

Claude Sonnet 4 vs Qwen3.5 4B

29 SHARED BENCHMARKS

Across 29 shared benchmarks, Claude Sonnet 4 scores higher on 23 and Qwen3.5 4B on 6. The widest gap is OmniScience Non-Hallucination, where Claude Sonnet 4 scores 70.9 against 13.4. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $3.00 per million input tokens).

ANTHROPICVSALIBABA29 SHARED23–6 HEAD-TO-HEAD

BenchmarkClaude Sonnet 4MARGINQwen3.5 4B

AA Agentic Index39.236.3

AA Intelligence2520

AA-LCR70.363

AA-Omniscience0.2-58.4

Artificial Analysis Coding Index37.622.6

critpt1.10

gdpval31.28.3

GPQA Diamond7877.1

HLE10.79.9

HMMT 202515.976.8

IFBench5559.2

ifeval87.689.8

LiveCodeBench v648.555.8

mmlu_redux93.688.8

MMLU-Pro8479.1

mmmlu86.576.1

MMMU74.473.4

MMMU-Pro62.465.4

multichallenge57.149

OmniScience Accuracy22.715.1

OmniScience Non-Hallucination70.913.4

scicode4018.3

supergpqa55.752.9

TauBench V3 - Banking16.76.8

Terminal-Bench 2.136.325.8

Terminal-Bench Hard31.118.2

τ²-Bench (Retail)7571.9

τ²-Bench Telecom (AA run)64.692.1

τ³-Bench13.88.2

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (anthropic-official, direct), otherwise the lowest tracked offer.

Compare Claude Sonnet 4 withALL PAIRINGS →

AnthropicClaude Opus 4.8 Claude Opus 4.6 Claude Sonnet 4.5 Claude 3.5 Sonnet Claude Fable 5 Claude Opus 4.5

Across the fieldGemini 3.6 Flash Nemotron 3 Super 120B A12B GPT-5.5 GPT-4o GPT-5.6 Sol Gemini 3.1 Pro GPT-5.4 GPT-5.2

Compare Qwen3.5 4B withALL PAIRINGS →

AlibabaQwen3.5 35B A3B Qwen3 VL 4B (Reasoning)Qwen3.5 397B A17B Qwen2.5 Instruct 72B Qwen3.5 27B Qwen3.6 Plus

Across the fieldClaude Sonnet 5 Muse Spark 1.1 GLM-5.1 GLM-5.2 o4-mini Claude Sonnet 4.6 o1 Gemma 4 31B