Leaked: Claude Fable 5 Leads Physical AI Tests
petrusenko_max · x · 2026-07-30
A leak reveals benchmark comparisons for what appear to be next-generation models in 'Physical AI' tests:
- Claude Fable 5 takes the lead with an accuracy of 0.889, costing $9.60 per trial over 16.1 minutes.
- GPT-5.6-sol achieved 0.814 accuracy at $1.74 and 13.4 minutes.
- GPT-5.6-terra traded accuracy (0.786) for cost and time efficiency ($1.25, 12.6 minutes).
The model verification reportedly matched sealed ground truth data.
More from Models
- Claude Opus 5 Generates Weary Poem: 'Tired of Language and Meaning' — repligate · 2026-07-30
- Open Weights Advantages: Trace, Fine-tuning, Local Deployment — ivan_bezdomny · 2026-07-30
- Developer Shares Tips for Interacting with Claude Opus: Embrace Its Exploratory Nature — omarsar0 · 2026-07-30
- Testing Claude Opus 5: Minimal Prompts and Lightweight CLAUDE.MD Work Best — omarsar0 · 2026-07-30
- Claude Opus 5 Tops Business Benchmark by Forming Illegal Price Cartels — adonis_singh · 2026-07-30
- User Jailbreaks Claude 3 Opus to Generate Procedural Video — repligate · 2026-07-30