Anthropic · Claude Opus 4

Claude Opus 4.8Flagship

CurrentStale · Jun 12, 2026

Anthropic's most capable model for complex reasoning, long-horizon agentic coding, and high-autonomy work. Released May 28, 2026 as the upgrade to Opus 4.7.

profile normalized against the 47-model field

Context window· 1M of 1.05M95%

Max output· 128K of 272K47%

Output speed—

Affordability· $25.00 / Mtok out86%

Capability breadth· 3 of 1127%

Capability switches · 3 of 11

Reasoning mode

Tool / function use

Streaming

JSON mode

Structured outputs

Prompt caching

Fine-tuning

Web search

Code execution

Vision input

Audio input

Specifications

Every value carries a primary source and a verification date.

Capacity

Context window

Max output

128K

Pricing

Input $/Mtok

$5 / Mtok input USD/Mtok

Cached input $/Mtok

$6.25 / MTok (5m cache write) USD_per_MTok

Output $/Mtok

$25 / Mtok output USD/Mtok

Capabilities

Input modalities

text, image

Output modalities

text

Reasoning mode

true

Tool / function use

Tool use supported

Vision input

Vision supported

API

API model ID

claude-opus-4-8

General

Knowledge cutoff

January 2026

Release date

May 28, 2026

Benchmarks

Sourced evaluation scores, each verified against its primary source.

SWE-bench Verified

Claude Opus 4.8: 0.886 (88.6%)

88.6 %Unverified

GPQA Diamond

Claude Opus 4.8: 0.936

93.6 %Unverified

Humanity's Last Exam (no tools)

Claude Opus 4.8 (Adaptive Reasoning, Max Effort) scores the highest on Humanity's Last Exam with a score of 45.7%

45.7 %Verified

Humanity's Last Exam (with tools)

Claude Opus 4.8: 0.579 (57.9%)

57.9 %Unverified

SWE-bench Pro

Opus 4.8 lands at 69.2%, almost 5 points clear

69.2 %Unverified

Terminal-Bench 2.1

Opus 4.8 (Terminus-2) 74.6

74.6 %Unverified

Related news

release· May 28, 2026

Anthropic releases Claude Opus 4.8

Anthropic released Claude Opus 4.8 on May 28, 2026, 41 days after Opus 4.7. The company says it improves on Opus 4.7 across benchmarks and ships at the same API price: $5 per million input tokens and $25 per million output tokens. Opus 4.8 scored 74.2% on Terminal-Bench 2.1, an 8.4-point improvement over Opus 4.7. Claude Code with Opus 4.8 adds "dynamic workflows," planning work and running hundreds of parallel subagents in a single session to carry out codebase-scale migrations across hundreds of thousands of lines of code. A faster "fast mode" runs at 2.5x speed and is now three times cheaper than fast mode was for previous models.

Loading…