
本期聚焦 GPT-5.6 的降价、Fast mode 与 Agent 审查降本,并梳理这些效率收益如何进入 Codex、ChatGPT Work 和 OpenAI 的推理栈。
major price cuts today:*80% drop for GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output *20% drop for GPT-5.6 Terra, to $2/$12 *GPT-5.6 Sol gets Fast mode in the API, up to 2.5x the speed for 2x the price, same intelligence https://t.co/erC6u4VoDR
今天有大幅降价:*GPT-5.6 Luna 降价 80%,现在每百万输入 token 0.20 美元、每百万输出 token 1.20 美元; *GPT-5.6 Terra 降价 20%,价格降至 2/12 美元; *GPT-5.6 Sol 在 API 中获得 Fast mode,速度最高是标准处理的 2.5 倍,价格为 2 倍,智能水平不变。
The work behind the new prices spans the model, inference stack, and agentic harness, from request routing and token generation to tool use and context management.Luna and Terra’s lower prices are also reflected in how usage is counted in Codex and ChatGPT Work, so your usage goes further.
新价格背后的工作覆盖模型、推理栈和 Agent harness,从请求路由、token 生成,到工具调用和上下文管理。Luna 和 Terra 的降价也会反映在 Codex 与 ChatGPT Work 的用量计算中,因此同样的用量可以走得更远。
We kept ourselves busy with GPT-5.6 Sol. It's pretty good at things.Rolling out
- 80% lower prices on Luna
- 20% lower prices on Terra
- faster speeds on GPT-5.6 Sol in /fast mode
- ~ 10X cheaper auto-approve mode ("review for me" in the app), which now uses Luna to prevent many high risk actions from the main agent
Towards intelligence too cheap to meter and many more breakthroughs ahead.
我们一直在忙着做 GPT-5.6 Sol。它在很多事情上表现得相当不错。正在推出:
- Luna 价格降低 80%;
- Terra 价格降低 20%;
- GPT-5.6 Sol 在 /fast mode 下速度更快;
- Auto-approve mode(应用里的「替我审查」)便宜约 10 倍;现在它使用 Luna,阻止主 Agent 执行许多高风险操作。
朝着「便宜到几乎不用计量的智能」继续前进,后面还有更多突破。
priority 请求继续有效。OpenAI Developers 还把它描述为「more intelligence per dollar」之外的另一种选择:在时间更重要的任务里,用预算换取更快完成,而不是换一个不同能力等级的模型。we've cut prices on luna by 80%, making it by far the most price-efficient model in its class.a lot of our research is about how to create incredibly efficient models for any given level of intelligence.excited to see what you all do with intelligence too cheap to meter!
我们把 Luna 的价格降了 80%,让它远远成为同档位中价格效率最高的模型。我们有很多研究都在解决同一个问题:怎样为每一个既定的智能水平打造极其高效的模型。很期待看到大家用「便宜到几乎不用计量的智能」做出什么。
With all due respect, things have evolved way beyond token in / token out. If tomorrow you wake up with your eyes on your back, I don’t think you would be very effective for the day.
恕我直言,事情已经远远超出了「输入 token、输出 token」的范围。假设你明天醒来时眼睛长在背后,我不认为你那天会有很高的工作效率。
this but for tibo resets - who's building it?
这个能不能也做一个「Tibo 重置」版?谁在做?
Inkling-Small is comparable to Inkling at a quarter the size. Weights are open, fine-tunable on Tinker today. Look forward to seeing what people make with it.
Inkling-Small 的能力可与 Inkling 相比,但规模只有四分之一。模型权重已开放,今天就可以在 Tinker 上微调。期待看到大家用它做出什么。
priority 请求继续有效。 原帖This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.