原文

Tibo@thsottiaux

2026-07-30 05:25· 35分钟前

跳到正文

AI 摘要

OpenAI 发布 GPT-5.6 Sol，部署后通过自我改进 GPU 内核与推测解码，实现服务成本降低 20%、token 生成效率提升 15%+。该模型在训练与推理全栈上持续优化自身效率。

Efficiency！

In two steps a） Train fantastic model b） Use fantastic model to make everything better， including its own infrastructure， inference stack， kernels， etc， etc

https://openai.com/index/gpt-5-6-frontier-intelligence-efficiency/

OpenAIAfter deployment, we applied GPT-5.6 Sol to advance the frontier of efficiency by making itself more efficient to run. The results: - 20% lower serving costs fr...

OpenAI 产品更新推理部署/工程

在 X 查看原推导出 Markdown

44导出 Markdown

Tibo · @thsottiaux · X·2026-07-30 05:25·35分钟前

在 X 看原推· x.com

AI 摘要

Efficiency！

In two steps a） Train fantastic model b） Use fantastic model to make everything better， including its own infrastructure， inference stack， kernels， etc， etc

https://openai.com/index/gpt-5-6-frontier-intelligence-efficiency/

OpenAIAfter deployment, we applied GPT-5.6 Sol to advance the frontier of efficiency by making itself more efficient to run. The results: - 20% lower serving costs fr...

OpenAI 产品更新推理部署/工程

Tibo@thsottiaux

2026-07-30 05:25· 35分钟前

跳到正文

AI 摘要

Efficiency！

In two steps a） Train fantastic model b） Use fantastic model to make everything better， including its own infrastructure， inference stack， kernels， etc， etc

https://openai.com/index/gpt-5-6-frontier-intelligence-efficiency/

OpenAIAfter deployment, we applied GPT-5.6 Sol to advance the frontier of efficiency by making itself more efficient to run. The results: - 20% lower serving costs fr...

OpenAI 产品更新推理部署/工程

在 X 查看原推导出 Markdown