THE HOOK
Perceptron's ex-Meta FAIR team just put egocentric multimodal reasoning on OpenRouter at pocket-change prices — and they made audio a first-class sensor.
THE GIST
Perceptron shipped Mk1.5 for drones, quadrupeds, and glasses at $0.15/$1.50 per M tokens on OpenRouter and AWS Marketplace. The model takes text/image/video/audio in a 36,864-token context, returns structured tracks, and claims 0.50 s P50 latency after a prior $21M Bessemer round.
THE TAKE
Pricing this low changes the economics of continuous onboard reasoning. Doppler-from-audio is a clever party trick; the real product is structured grounding cheap enough for a drone fleet. Vendor latency numbers always smile — make them prove it on your airframe before you retire the cloud uplink.
OPINION
Physical AI is splitting into brains (Perceptron, Skild) and runtimes (Isaac ROS); Mk1.5 undercuts Gemini Live / OpenAI realtime on cost, which is where industrial inspection lives. Hold them to zero-shot drone and quadruped control without platform fine-tunes — coverage via The Robotics Media.
- Story type: Breaking news
- Impact: Medium
- Confidence: Medium (multiple reports)
- Audience: Developers
- What changed: Perceptron released Mk1.5 on OpenRouter and AWS Marketplace with egocentric multimodal reasoning at $0.15/$1.50 per M tokens
- What to watch: Independent latency/quality benches vs Gemini Live and zero-shot drone/quadruped control demos
- Reading time: 1 minute
