VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

Claude Sonnet 4.5 vs Qwen3 VL 4B (Reasoning)

AnthropicvsAlibaba28 shared benchmarks262 head-to-head
BenchmarkClaude Sonnet 4.5Qwen3 VL 4B (Reasoning)
AA Agentic Index50.614.4
AA Intelligence37.47.7
AA-LCR68.323
AA-Omniscience-0.1-68.4
AI2D8784.9
AIME 20258874.5
Artificial Analysis Coding Index52.16.7
critpt1.10
gdpval40.413.8
GPQA Diamond83.464.1
HallusionBench59.964.1
HLE19.84.6
HMMT 202574.653.1
IFBench57.336.6
LiveCodeBench7151.3
LiveCodeBench v66451.3
MathVista79.879.5
mmlu_redux95.686
MMLU-Pro88.273.6
MMMU (val) (Pass@1)77.870.8
MMMU-Pro68.757
OmniScience Accuracy32.912.2
OmniScience Non-Hallucination50.98.2
OSWorld-Verified61.431.4
RealWorldQA70.373.2
scicode4517.1
Terminal-Bench Hard35.61.5
τ²-Bench Telecom (AA run)78.115.5

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.