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

Qwen3.6 27B vs Qwen3 VL 4B (Reasoning)

AlibabavsAlibaba32 shared benchmarks320 head-to-head
BenchmarkQwen3.6 27BQwen3 VL 4B (Reasoning)
AA Agentic Index27.514.4
AA Intelligence37.77.7
AA-LCR73.323
AA-Omniscience-20-68.4
Artificial Analysis Coding Index53.76.7
critpt1.10
EmbSpatial-Bench84.680.7
ERQA62.547.3
gdpval3213.8
GPQA Diamond87.864.1
HLE244.6
HMMT 202590.753.1
IFBench69.136.6
LiveCodeBench v683.951.3
mathvision85.160
MathVista87.479.5
mmlu_redux93.586
MMLU-Pro86.273.6
MMMU-Pro75.857
MMStar81.473.2
MV-Bench75.569.3
OCRBench89.481.6
OmniScience Accuracy19.612.2
OmniScience Non-Hallucination50.78.2
OSWorld-Verified63.931.4
RealWorldQA84.173.2
RefSpatial-Bench7045.3
scicode39.817.1
supergpqa6646.8
Terminal-Bench Hard34.81.5
VideoMMMU84.469.4
τ²-Bench Telecom (AA run)94.215.5

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.