Codesota · Models · Claude Opus 4.6Anthropic15 results · 6 benchmarks
Model card

Claude Opus 4.6.

Anthropicproprietary2 current SOTA

Imported from https://raw.githubusercontent.com/GAIR-NLP/AcademiClaw/main/README.md

§ 02 · Benchmarks

Every benchmark Claude Opus 4.6 has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01AcademiClawAgentic AI · Task agentsavg-score71.9%#1/62026-05-04source ↗
02AcademiClawAgentic AI · Task agentspass55.0%#1/62026-05-04source ↗
03HumanEvalComputer Code · Code Generationpass@196.3%#2/422026-01-01source ↗
04GPQA DiamondReasoning · Multi-step Reasoningaccuracy91.3%#2/74—source ↗
05React Native EvalsMobile Development · React Native Code Generationrequirement-satisfaction84.4%#2/10—source ↗
06React Native EvalsMobile Development · React Native Code Generationanimation-satisfaction77.4%#2/10—source ↗
07AcademiClawAgentic AI · Task agentssafety-score87.4%#3/62026-05-04source ↗
08AcademiClawAgentic AI · Task agentsavg-time-sec673.00#3/52026-05-04source ↗
09AcademiClawAgentic AI · Task agentstool-calls-per-task33.0%#3/62026-05-04source ↗
10AcademiClawAgentic AI · Task agentsavg-tokens-per-task-k1425.00#4/62026-05-04source ↗
11React Native EvalsMobile Development · React Native Code Generationasync-state-satisfaction84.6%#4/10—source ↗
12React Native EvalsMobile Development · React Native Code Generationnavigation-satisfaction93.3%#5/10—source ↗
13MMLUReasoning · Commonsense Reasoningaccuracy91.2%#7/642026-03-01source ↗
14HLEReasoning · Multi-step Reasoningaccuracy34.4%#12/74—source ↗
15HLEReasoning · Multi-step Reasoningaccuracy19.0%#40/74—source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area

Where Claude Opus 4.6 actually performs.

Agentic AI
1
benchmark
avg rank #2.5 · 2 SOTA
Computer Code
1
benchmark
avg rank #2.0
Mobile Development
1
benchmark
avg rank #3.3
Reasoning
3
benchmarks
avg rank #15.3
§ 04 · Papers

1 paper with results for Claude Opus 4.6.

  1. 2026-05-04· Agentic AI· 6 results

    AcademiClaw: When Students Set Challenges for AI Agents

    Junjie Yu, Pengrui Lu, Weiye Si, Hongliang Lu et al.
§ 05 · Related models

Other Anthropic models scored on Codesota.

Claude Opus 4
Undisclosed params · 14 results · 2 SOTA
Claude Sonnet 5
Undisclosed params · 2 results · 2 SOTA
Claude Sonnet 4
11 results · 1 SOTA
Claude Opus 4.5
4 results · 1 SOTA
Claude Opus 4.7
2 results · 1 SOTA
Claude Mythos Preview
1 result · 1 SOTA
Claude 3.5 Sonnet
Undisclosed params · 28 results
Claude Opus 4.5
Undisclosed params · 12 results
§ 06 · Sources & freshness

Where these numbers come from.

paper
6
results
Callstack Incubator
4
results
shadow-page-humaneval
1
result
anthropic-opus-4-6-announcement
1
result
codesota-shadow-mmlu
1
result
scale-hle-official
1
result
scale-ai-leaderboard
1
result
12 of 15 rows marked verified. · first result 2026-01-01, latest 2026-05-04.