Skip to content

Brief · 2 September 2026

What changed

Nori Robotics unveiled a $1,688 bimanual humanoid robot aimed at developers and research labs, expanding the sub‑$2k entry tier for mobile manipulation platforms. (source: norirobotics.com)

One number

1,688$

Price of Nori Robotics' bimanual humanoid, a sub‑$2k option for labs needing affordable mobile manipulation.

source ↗

Still vapor

Anthropic touts Claude Fable 5.1 as “up to 45 % cheaper for agentic work,” but the claim rests on internal pricing models and undisclosed workload mixes; independent cost‑per‑token benchmarks are still missing.

Today’s most concrete shift is the launch of Nori Robotics’ $1,688 bimanual mobile robot. At under $2 k it undercuts the nearest commercial humanoid platforms—Unitree’s G1 starts around $12 k and Boston Dynamics’ Atlas tops $150 k—making it viable for university labs, prototyping shops, and small‑scale startups. The robot offers two arms, 6‑DOF locomotion, and an open‑source SDK that integrates with ROS 2, letting engineers plug in existing perception stacks without bespoke firmware work. Early reviewers note the payload is modest (≈2 kg per arm) but sufficient for tabletop manipulation and basic research tasks.

Anthropic’s parallel announcement of Claude Fable 5.1 and Mythos 5.1 claims a 25‑30 % cost reduction for typical workloads and up to 45 % cheaper pricing for agentic use‑cases. The press release provides no third‑party cost analysis, and the pricing model hinges on opaque token‑pricing tiers that differ across regions. Operators should treat the savings claim as provisional until real‑world token‑cost data emerge.

Our catalog remains static—no new rigs verified in the last 30 days—so the Nori robot is the only fresh hardware entry. Teams planning to add embodied AI capability should compare the Nori platform’s price and open‑source stack against higher‑cost alternatives, while keeping an eye on Anthropic’s pricing rollout for downstream compute budgeting.

Composed by the MadCoolStuff editor pipeline · Groq · openai/gpt-oss-120b · 2026-09-02

Tags

What we read