Robocurve — Real-world evaluations of physical AI
Robocurve is a Public Benefit Corporation helping society understand the state of robot intelligence. We build open-source tools and independent benchmarks to evaluate how capable AI models and embodiments are in the real, physical world.
Backed byY Combinator We measure and report frontier robotics capabilities to the public. Opus 5 and the Tower of Hanoi $5 $10 $20 $50 $100 API cost per run (log) 0 20 40 60 80 100 score Opus (low) Opus (medium) Opus (high) Sol (high) Opus (low) Opus (medium) Opus (high) Sol (high) Jan '24 Jul '24 Jan '25 Jul '25 Jan '26 Jul '26 20 50 100 200 500 1,000 2,000 ×2.3/yr Claude 3 Haiku Celeris-1 ×7.8/yr o1 Mercury 2 ×3.4/yr GPT-5 Gemini 3.5 Flash-Lite ×2.1/mo Claude Fable 5 Gemini 3.7 Flash Intelligence level GPT-3.5 Turbo o1 GPT-5 Fable 5…
saved by
related reading
- BigFinanceBenchbigfinancebench.com
- CAIS AI Dashboarddashboard.safe.ai
- News — BenchCADbenchcad.com
- How Claude Performs on Robotics Tasks \ Anthropicanthropic.com
- Measuring the performance of our models on real-world tasks | OpenAIopenai.com
- Homeanthropic.com
- Cyborgism — LessWronglesswrong.com
- Physical Intelligence (π)pi.website
- ACT-1: A Robot Foundation Model Trained on Zero Robot Data | Sunday Robotics | The helpful robotics companysunday.ai
- GPT-6 Astra on robotic manipulationopenai.robocurve.org
- Frontier of General-Purpose Robotics (2025)adampatni.com
- TERMINAL-BENCHtbench.ai