Skip to content
Leaders Board.
Models/AnthropicClosed

Claude Opus 4.8

#33 of all models across the language-model boards.
#33
Overall rank
3.3
Podium score
0
Board wins
May 28, 2026
Released
·Chat·Reasoning·Coding·Agents and tools·Vision

Chat

Which model people prefer in blind head-to-head conversations.

18
Text Arena
1,481 · of 100
as “claude-opus-4-8-high”

Reasoning

Hard questions with checkable answers: science, math, puzzles, expert exams.

10
Epoch ECI
158.3 · of 100
as “Claude Opus 4.8”
17
ARC-AGI-2
72.1% · of 78
as “Claude Opus 4.8 (High)”
23
GPQA Diamond
91.0% · of 100
as “claude-opus-4-8_max”
26
LiveBench
76.2% · of 59
as “claude-opus-4-8-max-effort”

Coding

Fixing real repositories, building web apps, editing code.

25
WebDev Arena
1,555 · of 100
as “claude-opus-4-8-high”

Agents and tools

Calling functions and tools, using MCP servers, searching the web.

9
MCP Atlas
82.2% · of 34
as “claude-opus-4-8 (max)”
12
Search Arena
1,204 · of 33
as “claude-opus-4-8”

Vision

Understanding images, charts and documents.

9
Document Arena
1,475 · of 38
as “claude-opus-4-8-high”
16
Vision Arena
1,283 · of 100
as “claude-opus-4-8-high”

More from Anthropic

Claude Opus 5.5Claude Fable 5.1Claude Opus 5Claude Sonnet 5Claude Fable 5Claude Opus 4.7Claude Sonnet 4.6Claude Opus 4.6

Scores as published by each board on the capture date. Model names and logos belong to their owners; logos via logo.dev.

Weekly: the AI leaderboards with a new #1, Saturday mornings.