ClawForge - Partial Credit: leaderboard

ClawForge-Bench (2026) gives a command-line agent 362 generated workflow tasks over seeded calendars, inboxes, task boards, files and configuration, where pre-existing state may be partial, stale or duplicated, and grades the final state with executable validators. This board is the partial-credit score, the weighted mean of the evaluation-check scores for each task, averaged over tasks, from 0 to 100; higher is better.

Metric: Partial-credit score (0-100): weighted mean of a task's evaluation-check scores, averaged over tasks, times 100, on the ClawForge-Bench snapshot (17 generator-backed scenarios, 362 executable state-conflict workflow tasks over tasks, calendar, email, messaging, files and runtime configuration), command-line agent in multi mode with full interaction history and a 25-step budget, scored by result-first validators over the final normalized state; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-5.3 Codex82.38#124
2GPT-5.480.78#135
3O380.71#237
4GPT-5.278.83#185
5Claude Opus 4.678.68#113
6Claude Sonnet 4.674.73#173
7Kimi K2.574.04#266

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

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