Chat2Workflow - Education: leaderboard

Chat2Workflow (2026) tests whether language models build executable Dify workflows from multi-turn natural-language requests drawn from real business workflows, run against test cases. Resolve rate in percent on the education tasks; higher is better.

Metric: Resolve rate (%) on Chat2Workflow education tasks: share of test cases where the workflow the model generates for the current round of a multi-turn request, deployed on Dify 1.9.2, executes and its output satisfies every requirement of the instruction, judged by GPT-5.1; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Sonnet 4.560.61#261
2Gemini 3 Pro (Preview)59.6#119
3GPT-5.254.55#185
4GLM-4.752.53#380
5DeepSeek V3.132.32#511
6Qwen 3 235B A22B29.29#628
7GLM-4.625.25#498
8Qwen 3 32B23.23#972
9DeepSeek V3.222.22#419
10GPT-5.119.19#240
11Qwen 3 Coder 480B A35B Instruct17.17#652
12Kimi K2 (Thinking)15.15
13Kimi K213.13#443
14Qwen 3 8B0#1873
15Qwen 3 14B0#1269

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=chat2workflow-education · How It Works · Data refreshed daily, snapshot 2026-11-21.