Benchmarks / Agentic
ApprenticeBench CUA (NeoCognition)
A computer-use agent operates the Odoo ERP through its screens across 100 sequential accounts-payable tasks with diminishing mentoring. Maintained by NeoCognition. Score: Cumulative success rate.
ApprenticeBench CUA (NeoCognition)
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Fable 5.1 | max | 72.0% |
| 2 | GPT-6 Astra Codex | max | 68.0% |
| 3 | Fable 5 | xhigh | 43.0% |
| 4 | Opus 5 | max | 36.0% |
| 5 | GPT-5.6 Sol Codex | max | 26.0% |
| 6 | Gemini 3.8 Flash | max | 24.0% |
| 7 | GPT-5.5 Codex | xhigh | 20.0% |
| 8 | Muse Spark 1.3 | max | 19.0% |
| 9 | Kimi K3 | max | 18.0% |
| 10 | Sonnet 5 | max | 16.0% |
| 11 | GPT-5.6 Terra Codex | max | 16.0% |
| 12 | Gemini 3.7 Flash | max | 16.0% |
| 13 | Grok 4.6 | max | 13.0% |
| 14 | GPT-5.4 Codex | xhigh | 11.0% |
| 15 | Opus 4.7 | max | 7.0% |
| 16 | GPT-5.6 Luna Codex | max | 7.0% |
| 17 | Opus 4.6 | max | 5.0% |
| 18 | Sonnet 4.6 | max | 2.0% |