Two New Robotics Benchmarks Released: MicroPlanBench and MicroAnnotationBench First Rankings
chris_j_paxton · x · 2026-10-20
The team released two new robotics benchmarks. MicroPlanBench (next-action prediction) is led by GPT-6 Astra, with GPT-6.1 Sol close behind at 1/5 the cost. MicroAnnotationBench, which provides labels plus timing on long real-world clips, is topped by their own nanoannot-1, beating Gemini 4 Argon at 1/10 the cost. Cost efficiency is the standout finding of both leaderboards.
More from Research
- Single neuron steering makes Qwen3-8B insist it's an owl — Sauers_ · 2026-10-20
- Q-CTRL: over 99.5% of qubits sit idle in one fault-tolerant Shor's algorithm blueprint — MJBiercuk · 2026-10-20
- No Siegel zeros proved in just 3 pages, echoing the polynomial method's history — soumitrashukla9 · 2026-10-20
- Logarithmic returns come from opportunities spread across orders of magnitude of cost — DimitrisPapail · 2026-10-20
- Emergent curriculum without labels: prompt weights diverge 1.6x vs 0.75x over training — shangbinfeng · 2026-10-20
- Stackelberg Alignment turns LLM disagreement into a training curriculum via bandit leader — shangbinfeng · 2026-10-20