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

DeepSeek-R1 vs Qwen3.5 27B

DeepSeekvsAlibaba33 shared benchmarks627 head-to-head
BenchmarkDeepSeek-R1Qwen3.5 27B
AA Agentic Index3.154.6
AA Intelligence18.634.6
AA-LCR5672.3
AA-Omniscience-31.3-44
Artificial Analysis Coding Index24.634.9
browsecomp8.961
browsecomp_zh35.762.1
C-Eval91.890.5
critpt0.60.9
gdpval1.533
GPQA Diamond8185.8
HLE17.748.5
HMMT 202579.489.8
hmmt_feb_202576.792
IFBench3976.5
ifeval83.395
longbench_v258.360.6
mmlu_prox75.582.2
mmlu_redux93.493.2
MMLU-Pro8586.1
multichallenge4560.8
OmniScience Accuracy30.720.7
OmniScience Non-Hallucination10.524.9
scicode35.739.5
SWE-bench Multilingual30.569.3
SWE-bench Verified57.675
Terminal-Bench Hard6.132.6
vectara_answer_rate9799.8
vectara_avg_summary_length93.594.4
vectara_factual_consistency88.787.9
vectara_hallucination_rate11.312.1
τ²-Bench Telecom (AA run)11.493.9
τ³-Bench6.468.4

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.