Claude Opus 4.6

Released

Claude Opus 4.6 is an AI model released by Anthropic on Thursday, Feb 5 2026, 73 days after Claude Opus 4.5. Benchmark results (shown below) cover BullshitBench v2, ProgramBench, SWE-Bench Verified, Next.js Evals, BrowseComp, Humanity's Last Exam, and 2 more.

API pricing

Input
$5.00
Cached input
$0.50
5 min cache write
$6.25
1 hr cache write
$10.00
Output
$25.00
  • Reasoning or thinking is supported.
USD per 1M tokens

Available from

Amazon Bedrock$5.00$25.001M
Azureglobal$5.00$25.001M
Googleglobal$5.00$25.001M
Googleeu$5.50$27.501M
USD per 1M tokens

Benchmarks

Coding

ProgramBench
#5 of 17
SWE-Bench Verified
80.8%
#9 of 57
Next.js Evals
58%
#19 of 28

Agentic & tool use

BrowseComp
83.7%
#14 of 28
OSWorld-Verified
72.7%
#17 of 28

Reasoning & science

Humanity's Last Exam
53%
with tools
#17 of 42
GPQA Diamond
91.3%
#12 of 59

Robustness

BullshitBench v2
87%
#5 of 77

Source: ProgramBench, retrieved 11 September 2026 · Source: BenchLM, retrieved 31 August 2026. Every other score here is the figure the lab published at launch.

Compare Claude Opus 4.6 with

Claude Opus 4.6

Suggested comparisons

Frequently asked questions

Claude Opus 4.6 was released by Anthropic on Thursday, Feb 5 2026.

All Anthropic releases

27 tracked