The StreamlinersStrict ZDR Verified

Particle AI

Token-efficient coding and agentic workflows.

Laboratory Overview

Particle AI is an inference and post-training lab dedicated to radically driving down the cost per unit of intelligence. Creators of Particle-1, Particle AI develops reinforcement learning environments that train frontier open-weight models to solve software engineering and agentic tasks with substantially fewer reasoning and tool-call tokens. On ARMES, all Particle AI requests run via verified Zero Data Retention enterprise pipelines.

Core Research Focus:Token-efficient agentic codingTerminal and shell CLI workflowsCodebase refactoring and patch generationData engineering and pipeline scriptingLow-overhead tool-calling loopsReduced-token reasoning minimization

Privacy & Zero Data Retention

Strict Zero Data Retention (ZDR): Requests are processed ephemerally in RAM. No user data, prompts, or completions are stored, indexed, or used for model training.

No user telemetry sent to Particle AI

Particle AI Foundation Models (1)

Available under ARMES unified subscription without separate API keys or billing agreements.

Tier:

Sparse Mixture-of-Experts (MoE) transformer, adapter-post-trained (rank-64 LoRA, ~1.25% of parameters) on DeepSeek V4.1 Flash

MoEFree

Particle-1

Particle-1 is Particle AI's first model (released October 28, 2026): a rank-64 LoRA post-train of DeepSeek V4.1 Flash (~1.25% of parameters) that teaches the model to reach correct agentic-coding answers with 28.4%–52.2% fewer output tokens at roughly accuracy parity, priced at $0.05/$0.50 per 1M tokens with a 1,048,576-token context and 131,072-token output budget.

Output Token Reduction:28.4%–52.2%

1.05M ctx

$0.05 / $0.50 /MTok

Intelligence Dossier Try in Chat

Explore Other AI Labs on ARMES

Compare models across the 27+ leading laboratories unified in your workspace.