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 Opus 5

AnthropicvsAnthropic30 shared benchmarks228 head-to-head
BenchmarkClaude Sonnet 4.5Opus 5
AA Agentic Index50.659.2
AA Intelligence37.463.1
AA-LCR68.378.7
AA-Omniscience-0.137.1
arc_agi_163.797.5
ARC-AGI-213.690.4
arena_elo14561487
arena_text_factuality14481481
Artificial Analysis Coding Index52.178
browsecomp24.190.8
coding_arena_elo13921662
critpt1.129.1
gdpval40.467.2
GPQA Diamond83.493.7
HealthBench44.267.1
HLE19.864.7
HLE (with tools)3264.7
MCP Atlas59.585.8
MMMU-Pro68.784.7
OmniScience Accuracy32.960.9
OmniScience Non-Hallucination50.940.5
OSWorld-Verified61.44
scicode4555.7
simplebench54.380.6
simpleqa_verified23.656.7
SWE-bench Multilingual6889.5
SWE-bench Pro43.679.2
SWE-bench Verified8296
TauBench V3 - Banking24.544.7
Terminal-Bench 2.155.889.1

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.