Particle AI releases Particle-1, a 1.0M-context agentic coding model priced at $0.25/$1 per 1M tokens

Particle AI released Particle-1 on October 28, 2026, a token-efficient agentic coding model post-trained on DeepSeek V4.1 Flash. It is listed on OpenRouter with a 1.0M-token context window at $0.25 per 1M input tokens and $1 per 1M output tokens.

Key specifications

• Developer: Particle AI
• Model ID: particle-ai/particle-1
• Base model: DeepSeek V4.1 Flash (post-trained)
• Context window: 1.0M tokens
• Pricing: $0.25 per 1M input tokens, $1 per 1M output tokens
• Release date: October 28, 2026
• Parameter count, training cutoff, benchmark scores: not yet disclosed

Intended use

According to its OpenRouter listing, Particle-1 is suited for software engineering work such as terminal workflows, codebase changes, and data engineering tasks. The listing says the model is trained to complete these tasks with substantially fewer reasoning and tool-call tokens while keeping accuracy close to its base model. That efficiency and accuracy claim comes from the listing, and no benchmark figures accompany it.

Availability

Particle-1 is accessible through OpenRouter. The listing states it is not currently available through EU in-region routing.

What this means

Particle-1 targets the cost of agentic coding workloads, where long reasoning traces and repeated tool calls drive token spend. Its stated approach is to cut those tokens rather than to raise raw capability, with accuracy held close to DeepSeek V4.1 Flash. Because the listing gives no benchmark results, the size of the efficiency gain and the accuracy trade-off remain to be measured by independent testing. Teams considering it should note the lack of EU in-region routing.