Particle AI
About Particle AI
AI lab training token-efficient models for agentic coding and data engineering, served through the Particle Inference API.
Performance
Recent Gateway performance across models from Particle AI.
Throughput
—
No recent data
Time to first token
—
No recent data
Weekly usage
—
No recent data
Performance metrics will appear after this lab’s models serve Gateway traffic.
Latest models
Recent releases with availability, context, pricing, and usage details.
Particle-1 is a token-efficient agentic coding model from Particle AI, post-trained from DeepSeek V4.1 Flash with reinforcement learning to solve tasks with fewer reasoning and tool-call tokens. Particle reports about 45% fewer tokens on average at accuracy within the base model's confidence intervals.
Input
Output
28 Oct 2026