Particle-1, launched by the Particle Team, reduces output tokens by 45%.

iconKuCoinFlash
Share
AI summary iconSummary
On-chain news: The Particle team has launched Particle-1, a model based on DeepSeek V4.1 Flash, designed to enhance token efficiency. On software engineering tasks, Particle-1 reduces output tokens by 45%, with an average cost of $0.2211 per task—lower than DeepSeek V4.1 Flash ($0.3368) and GLM 5.3 Flash ($0.6523). OpenClaw users must update to v2026.10.10 to access the model, which supports standard inference and zero data retention. Additional token listings may follow as adoption increases.
ME AI message: The Particle team has launched Particle-1, fine-tuned on DeepSeek V4.1 Flash, with the core goal of improving token efficiency. On software engineering tasks, the model reduces average output tokens by approximately 45%, with a cost of $0.2211 per task—lower than DeepSeek V4.1 Flash at $0.3368 and GLM 5.3 Flash at $0.6523; the average output tokens for the three models are 59,580, 83,246, and 98,155 respectively, with task pass rates of 72.57%, 71.15%, and 66.82%. On Terminal-Bench 2.1, ReactBench, ADE Bench, and DeepSWE 1.1, Particle-1 uses 28.4%–52.2% fewer output tokens than the base model, with Pass@1 within the base model’s confidence interval. Training employed Cascade RL, rank-64 LoRA (accounting for ~1.25% of parameters), GSPO, and the Miles framework. OpenClaw users must update to v2026.10.10 to use it; standard inference retains zero data. (Source: InFoQ)
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.