OmniaBench - DAG-S: leaderboard

Metric: Pass@1 (%; 200 single-turn tasks refined from the DAG route; a task counts as solved only when every rubric item is satisfied (Program tasks by a VerifyCode check); up to 200 tool-call steps; challenging set of 644 tasks evaluated June 29 to July 15, 2026). Source: arxiv.org. Saturation forecast: Around December 2026. 22 models tracked.

Top models

#ModelScore
1GLM-5.2 (Max)69
2GPT-5.5 (High)64.5
3DeepSeek V4 Pro (Max)63.5
4Claude Sonnet 5 (Max)60.5
5GPT-5.6 Sol (High)59
6Kimi K2.6 (Thinking)57.5
7Claude Opus 4.7 (Max)57.5
8DeepSeek V4 Flash (Max)51
9Qwen 3.7 Max (Thinking)48.5
10GLM-5 (Thinking)38.5
11Gemini 3.5 Flash (High)35.5
12Qwen 3.6 27B35
13Qwen 3.6 35B A3B32
14Gemini 3.1 Pro (Preview)29
15Qwen 3.5 397B A17B26

Interactive version: theaggregate.ai/benchmark?slug=omniabench-dag-s · How It Works · Data refreshed daily, snapshot 2026-09-29.