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

Llama 3.1 Instruct 405B vs Qwen3.5 397B A17B

MetavsAlibaba24 shared benchmarks321 head-to-head
BenchmarkLlama 3.1 Instruct 405BQwen3.5 397B A17B
AA Agentic Index6.353.3
AA Intelligence8.534.3
AA-LCR24.772.7
AA-Omniscience-17.1-30.8
Artificial Analysis Coding Index14.548.2
C-Eval72.593
critpt01.7
gdpval035.8
GPQA Diamond51.589.3
GSM8K96.866.7
HLE4.229
IFBench3978.8
ifeval88.692.6
longbench_v236.163.2
mmlu_redux86.294.9
MMLU-Pro73.487.8
mmmlu73.888.5
OmniScience Accuracy23.230.8
OmniScience Non-Hallucination4917.3
scicode29.942
SWE-bench Pro11.250.9
SWE-bench Verified24.576.4
Terminal-Bench Hard6.840.9
τ²-Bench Telecom (AA run)1995.6

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.