Pinned
Today we're releasing Perceptron Mk1: frontier video and embodied reasoning.
00:00
Perceptron AI
57 posts
Perceptron AI
@perceptroninc
The layer of intelligence that bridges the physical and digital worlds.
perceptron.inc
Joined September 2024
- 1/ Introducing Isaac 0.1 — our first perceptive-language model. 2B params, open weights. Matches or beats models significantly larger on core perception. We are pushing the efficient frontier for physical AI. perceptron.inc/blog/introduci…
- Today we’re releasing the technical details behind Isaac 0.1. We built Isaac to demonstrate that with the right principles, a simple recipe can reach competitive performance with a small model. Check out the report at: github.com/perceptron-ai-…
- We are Perceptron AI, a new foundation model company from @ArmenAgha, @AkshatS07. Foundation models transformed the digital realm, now it’s time for the physical world. We’re building the first foundation models designed for real-time, multi-modal intelligence across the real
- Designing multimodal systems is challenging. To build with the community we’re sharing the design behind TensorStream - a convenient tensor-like interface for interleaved multimodal data. Internally our training and inference code is built on top of this.
- Perceptron’s platform is here — built for Physical AI Developers can now use Isaac-0.1 or Qwen3VL 235B via: Perceptron API — fast, reliable multimodal intelligence Python SDK — simple, grounded prompting for vision + language Build apps that see and understand the world.
00:00 - Replying to @perceptroninc9/ Demo perceptron.inc/demo Open Weights 🤗 huggingface.co/PerceptronAI SDK github.com/perceptron-ai-… Try on fal blog.fal.ai/isaac-the-firs…
- Replying to @perceptroninc2/ Why it matters: Isaac is tuned for where the physical world needs intelligence. Capability alone isn’t enough—you need capability at cost, power, and tail-latency constraints.
- Replying to @perceptroninc3/ Performance: We’re releasing our base and post-trained 2B checkpoints. Isaac leads its size class on grounding/localization & VQA; competitive with much larger systems. Details will be available soon through our technical report.
- Replying to @perceptroninc4/ In-context learning for perception: show a few annotated examples (defects, safety flags, etc.) and the model adapts in-prompt. No YOLO-style fine-tuning or custom detector stacks.
- Replying to @perceptroninc5/ Grounded spatial intelligence: precise pointing & localization through occlusion/clutter. Ask “what’s broken in this machine?” → grounded answer with highlighted regions.
- Replying to @perceptroninc7/ Conversational Pointing: language and vision stay in lockstep - every claim visually cited. A new surface area for visual reasoning.
- Replying to @perceptroninc10/ Join us to push this frontier: [email protected].
- Replying to @perceptroninc6/ Visual Reasoning: OCR, dense scenes, and cluttered layouts across resolutions with dynamic image resolution natively supported.
