WorkBuddy Bench (CodeBuddy Code) - Web: leaderboard
Tencent WorkBuddy Bench (2026) evaluates coding agents on colloquial, role-played requests reverse-engineered from real work. This board is the Web track, 70 front-end and GUI tasks scored from 0 to 100 by rules, LLM or VLM judges and agent rubrics, with each model run in the CodeBuddy Code harness in think mode, mean of three runs; higher is better.
Metric: Score (0-100) on the 70 front-end and GUI tasks scored by rule, LLM or VLM and agent rubrics, run in the CodeBuddy Code harness in think mode, mean of three runs; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | CodeBuddy Code + Claude Opus 4.8 (Thinking) | 68.14 | |
| 2 | CodeBuddy Code + Hy3 | 67.71 | |
| 3 | CodeBuddy Code + GLM-5.2 | 67.43 | |
| 4 | CodeBuddy Code + GPT-5.5 | 61.14 | |
| 5 | CodeBuddy Code + MiniMax-M3 | 58 | |
| 6 | CodeBuddy Code + DeepSeek-V4-Pro (Thinking) | 54.57 | |
| 7 | CodeBuddy Code + DeepSeek-V4-Flash (Thinking) | 47.29 |
Interactive version: theaggregate.ai/benchmark?slug=workbuddy-bench-codebuddy-code-web · How It Works · Data refreshed daily, snapshot 2026-11-13.