Claude Opus 4.6 benchmark ranking on every AI leaderboard

Name: Claude Opus 4.6
Author: Anthropic

closed weights · #7 overall, podium score 22.7 · top 10 on 7 of 14 · released Feb 5, 2026

Anthropic

14 of 14 boards

Rank	Board	Category	Score	Of
2	Text ArenaChat · as “claude-opus-4-6-high”	Chat	1,505	100
2	Search ArenaAgents and tools · as “claude-opus-4-6-search”	Agents and tools	1,253	33
3	Document ArenaVision · as “claude-opus-4-6-high”	Vision	1,507	38
4	Vision ArenaVision · as “claude-opus-4-6-high”	Vision	1,299	100
10	HLEReasoning · as “claude-opus-4-6-thinking-max”	Reasoning	34.4%	42
16	MCP AtlasAgents and tools · as “claude-opus-4-6 (max)”	Agents and tools	76.8%	34
17	MultiChallengeChat · as “claude-opus-4-6 (Non-Thinking)”	Chat	56.0%	29
19	ARC-AGI-2Reasoning · as “Claude Opus 4.6 (120K, High)”	Reasoning	69.2%	78
25	Epoch ECIReasoning · as “Claude Opus 4.6”	Reasoning	155.4	100
27	WebDev ArenaCoding · as “claude-opus-4-6-high”	Coding	1,547	100
30	GPQA DiamondReasoning · as “claude-opus-4-6_32K”	Reasoning	90.5%	100
37	LiveBenchReasoning · as “claude-opus-4-6-thinking-auto-high-effort”	Reasoning	74.5%	59
4	MMLU-ProReasoning · stale since 2026-03-11 · as “Claude-4.6-Opus(Thinking)”	Reasoning	89.1%	100
4	SWE-bench VerifiedCoding · stale since 2026-02-26 · as “Claude 4.6 Opus”	Coding	75.6%	39

Category ranks and more models

Rank in each category

More from Anthropic

Sources: each leaderboard's own published results page, captured every day at 04:30 UTC. Scores exactly as each board published them; we run no evaluations. Logos via logo.dev; trademarks belong to their owners.

Rank in each category

More from Anthropic

New #1s by email