Skip to content
Leaderboard / Anthropic

Claude Sonnet 4.5

Anthropic 2025-09-29 CLOSED Proprietary
Other settings:High, 32K, 2025-09-29 · 65
Verdict

Claude Sonnet 4.5 is a solid mid-pack model: #163 of 406 by SIscope score. Strongest at following instructions, weaker at math. On the pricey side: costs more than 86% of priced models.

SIscope score
54
#163 of 406
Human-vote rank
#79
of 419 · score 1439
Price per 1M tokens, blended
$6
$3 in · $15 out
Context window
1M
1,000,000 tokens

Shape

Each spoke is a category, 0 at the centre, 100 at the edge.

CodingAgentsno data yetReasoningMathChatInstructions

Category scores

Percentiles: 90 means better than about 90% of tested models.

Coding49
Agents & toolsNo data yet–
Reasoning & knowledge51
Math48
Chat quality81
Following instructions90

What goes into each category · Compare with DeepSeek V4.1 Flash

Human-vote score over time

Text arena, recent updates.

143014381445Oct 20: 1438.5 (rank #78)Oct 22: 1438.4 (rank #79)Oct 28: 1438.5 (rank #79)Oct 20Oct 28

Coding votes over time

WebDev arena, recent updates.

138013851390Oct 16: 1385.1 (rank #100)Oct 18: 1385.2 (rank #101)Oct 19: 1385.1 (rank #104)Oct 22: 1385.3 (rank #105)Oct 23: 1385.3 (rank #106)Oct 27: 1385.3 (rank #107)Oct 28: 1384.9 (rank #108)Oct 16Oct 28

Independent test runs

Run by Epoch AI on the same harness for every model. Bars show the share of questions answered correctly.

GPQA Diamond74%
PhD-level science multiple choice questions
Better than 46% of 333 tested models. Run Sep 29, 2025.
SimpleQA Verified24%
Short factual questions; rewards not making things up
Better than 19% of 100 tested models. Run Aug 10, 2026.
OTIS Mock AIME 2024-2536%
Competition maths problems in the style of AIME
Better than 26% of 313 tested models. Run Sep 29, 2025.
SWE-bench Verified71%
Fixing real GitHub issues in Python repositories
Better than 29% of 35 tested models. Run Feb 5, 2026.

Human-vote arenas

People compare two anonymous answers and pick the better one. Scores work like chess ratings.

ArenaRankScoreRange (95%)VotesUpdated
Text arena
Everyday chat, blind human votes
#79 of 41914391436–144182,621Oct 28, 2026
WebDev arena
Building web apps, blind human votes
#108 of 15213851378–139215,679Oct 28, 2026
Vision arena
Questions about images, blind human votes
Not on this board yet
Agent arena
Real multi-step agent sessions
Not on this board yet

Source: Arena (formerly LMArena) leaderboard dataset, CC BY 4.0. Fetched Nov 4, 2026.

Price

Input
$3
per 1M tokens
Output
$15
per 1M tokens
Blended
$6
3 in : 1 out, per 1M
Context window
1M
max tokens per request
Speed
Coming soon
our own tests

List price from the OpenRouter models API (anthropic/claude-sonnet-4.5), Nov 4, 2026. Providers can charge differently.