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-V4-Pro vs Gemini 2.5 Pro

DeepSeekvsGoogle35 shared benchmarks296 head-to-head
BenchmarkDeepSeek-V4-ProGemini 2.5 Pro
AA Agentic Index63.37.2
AA Intelligence5327
AA-LCR7066
AA-Omniscience-10.6-14.3
arena_elo14571446
arena_text_factuality14501462
Artificial Analysis Coding Index59.446.7
browsecomp83.49.9
coding_arena_elo15821224
critpt132.6
gdpval498.5
Global-MMLU-Lite89.388.6
GPQA Diamond90.584.4
HLE48.222.5
HLE (with tools)48.228.4
hmmt_nov_202594.480
IFBench76.549
LiveCodeBench93.582.7
MMLU-Pro87.586
OmniScience Accuracy4339
OmniScience Non-Hallucination12.212.6
scicode5043
simplebench50.962.4
simpleqa57.950.8
simpleqa_verified57.956
SWE-bench Verified80.667.2
TauBench V3 - Banking30.19.7
Terminal-Bench 2.172.128.5
Terminal-Bench Hard46.226.5
vectara_answer_rate97.299.1
vectara_avg_summary_length153.8106.4
vectara_factual_consistency91.493
vectara_hallucination_rate8.67
τ²-Bench Telecom (AA run)96.254.1
τ³-Bench25.89.3

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.