CoworkOS - JobBench: leaderboard
CoworkOS paper comparison of OpenHands, Codex, DeepSeek Harness and CoworkOS, all using GLM-5.3 High effort, 256K context and64K maximum output. Harness identities remain separate. Each benchmark uses its native scoring protocol; domain averages equally weight constituent benchmarks. GDPVal Elo uses all220 scheduled tasks, invalid/timeouts as losses, baseline1000; Productivity average substitutes exact-count tie-adjusted win rate for Elo. JobBench. Main split65 tasks,7200-second timeout; mean LLM-judged deliverable score.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.3 (high) (CoworkOS Run: harness=CoworkOS; context=256K; output=64K) | 59.88 | |
| 2 | GLM-5.3 (high) (CoworkOS Run: harness=DeepSeek%20Harness; context=256K; output=64K) | 56.15 | |
| 3 | GLM-5.3 (high) (CoworkOS Run: harness=OpenHands; context=256K; output=64K) | 55.02 | |
| 4 | GLM-5.3 (high) (CoworkOS Run: harness=Codex; context=256K; output=64K) | 23.42 |
Interactive version: theaggregate.ai/benchmark?slug=coworkos-jobbench · How It Works · Data refreshed daily, snapshot 2026-11-05.