The new AI assistant from Particle is now accessible through the OpenRouter platform as part of a new partnership to deliver a streamlined inference service. The tool, known as Particle-1, is a coding model post-trained on DeepSeek V4.1 Flash and is engineered to reduce the resource load of AI tasks, cutting output token usage by 28% to 52% across four coding benchmarks.
Particle says the model’s average token reduction of approximately 45% will lower costs and completion times for autonomous agent tasks while maintaining accuracy levels near those of the base model. Particle-1 costs $0.25 per million input tokens and $1 per million output tokens, and it features a 1.05M context window.