Anthropic's Fable 5.2, utilizing a basic model, shows substantially improved skills in controlling physical robots, even on un-trained tasks. It outperformed other models and most AI systems on the RoboDojo benchmark, which features a simulated two-arm robot.
Anthropic Fable 5.2 Improves Robot Control, Outperforms on BenchmarkAI News
Anthropic's Fable 5.2, utilizing a basic model, shows substantially improved skills in controlling physical robots, even on un-trained tasks. It outperformed other models and most AI systems on the RoboDojo benchmark, which features a simulated two-arm robot.