← Back to Model Leaderboard

Claude Sonnet 4.6

Official information

Release date: Feb 17, 2026
Official model ID: claude-sonnet-4-6

Vendor source: Claude Sonnet 4.6
Identity verified: Aug 13, 2026

API pricing

Verification: Verified
Verified at: Oct 27, 2026, 06:10:02 GMT+8
Value benchmark tier: claude-sonnet-4-6-standard

Input / output per million tokens
Tier / region	Input / output · per MTok	Context pricing tier	Billing	Validity	Source
claude-sonnet-4-6-standard global	$3 / $15	No upper pricing tier	Standard online · Cache miss All day	Current	Official rates Oct 27, 2026, 06:10:02 GMT+8

Rankings by category

Overall

#30High 0.91

43.8

Overall — Source breakdown
Source	Percentile	Raw score	Config folding
Epoch AI	39.8	152.8	Single config
LiveBench	33.7	73.4	Single config
LMArena	57.8	1457.8	Single config

Cross-source percentile spread 24.1

Coding

#38Medium 0.68

33.8

Coding — Source breakdown
Source	Percentile	Raw score	Config folding
DeepSWE	6.3	29.9	Single config
LiveBench	29.3	57.3	Single config
LMArenatext-coding	81.4	1503.3	Single config
· webdev	50.0	1522.7	Single config

Cross-source percentile spread 59.4

Agent

#37Medium 0.57

6.3

Agent — Source breakdown
Source	Percentile	Raw score	Config folding
DeepSWE	6.3	29.9	Single config

Reasoning

#38High 0.83

36.2

Reasoning — Source breakdown
Source	Percentile	Raw score	Config folding
Epoch AI	10.4	0.8	Median of 4 configs
LiveBench	42.4	85.9	Single config
LMArenatext-math	42.2	1462.5	Single config
· text-hard	69.6	1484.5	Single config

Cross-source percentile spread 45.5

Value

#30High 0.91

31.2

Aliases & configurations

claude-sonnet-4-6
claude-sonnet-4-6-thinking-auto-medium-effort· thinking auto / medium effort

Methodology & data sources