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 Opus 4.5 vs Opus 5

AnthropicvsAnthropic34 shared benchmarks232 head-to-head
BenchmarkClaude Opus 4.5Opus 5
AA Agentic Index59.659.2
AA Intelligence41.963.1
AA-LCR7678.7
AA-Omniscience1437.1
arc_agi_18297.5
ARC-AGI-237.690.4
arena_elo14731487
arena_text_factuality14581481
Artificial Analysis Coding Index47.878
browsecomp3790.8
coding_arena_elo14951662
critpt4.629.1
gdpval47.367.2
GPQA Diamond8793.7
HLE30.864.7
HLE (with tools)43.464.7
livebench_agentic_coding39.765.2
livebench_coding79.781.5
livebench_data_analysis74.474.5
livebench_instruction_following62.563.8
livebench_language81.388.7
livebench_math90.495.7
livebench_reasoning80.191.2
MCP Atlas69.885.8
MMMU-Pro7484.7
OmniScience Accuracy46.660.9
OmniScience Non-Hallucination3940.5
OSWorld-Verified66.34
scicode5055.7
simplebench6280.6
simpleqa_verified41.856.7
SWE-bench Multilingual77.589.5
SWE-bench Pro57.179.2
SWE-bench Verified81.596

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.