Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Yesterday was Grok 4.7 (pelicans) and MiMo v2.6 Flash/Pro (…
Marked important, but published more than 30 days ago, so it moved down to this band.
On this page12 sections
The editors' notes below — why it matters, the context, the learning path — are written in Chinese.
Record
The signal
#Yesterday was Grok 4.7 (pelicans) and MiMo v2.6 Flash/Pro (more pelicans). Today Anthropic released Claude Opus 5.5, and around an hour later OpenAI released GPT-6 Sol and GPT-6 Luna. It's going to take a while to get a good read on all of these new models, but here are my impressions so far. GPT-6 Sol and Luna are half the price of their GPT-5.6 equivalents GPT-5.6 Luna was already my favorite model for building applications against, because it combined excellent performance with being really cheap. Somehow GPT-6 Luna is half the price of that again - and GPT-6 Sol had a similar reduction compared to GPT-5.6 Sol. Here's what the pricing landscape looks like today: Model Input Cached input Output GPT-6 Luna $0.10/M $0.01/M $0.50/M GPT-5.6 Luna $0.20/M $0.02/M $1.20/M Grok 4.7 $2/M $0.50/M $6/M GPT-6 Sol $2/M $0.20/M $10/M GPT-5.6 Terra</…
Why now
Why it matters
#两家实验室在同一天把次顶档模型的单价砍到一半上下，最顶档却没动。对大多数团队，这意味着换代的成本账要重算——但决定账单的不是标价，而是缓存命中率和推理档位；而这一轮里恰好有一个把推理档开到最高、结果什么都没返回的反例。
Background
Technical context
#API 价格通常分输入、缓存输入、输出三项。智能体每轮都会重发完整上下文，所以长对话里绝大多数输入 token 命中缓存，缓存价往往比标价更决定成本。推理模型另有推理档位（effort），档位越高思考 token 越多，而思考 token 按输出计费，且受同一个输出上限约束。
Audience
Who should care
#Next
Learning path
#- 先读《旗舰模型发布解读与换代判断》：把公告里的数字分成可直接比、换算后可比、基本不可比三类
- 把这张价格表按你的实际用量重算一遍：缓存输入占你输入 token 的多少
- 读本站 8 月的 Qwen3.8 默认 xhigh 实测，与这里的 Opus 5.5 max 并排看
- 为每个推理档位设一个输出上限，并在自己的任务上量一次平均思考长度
- 等 Sonnet 5.5 与 Haiku 5.5 出来后，再看低端的价格是否跟进
Network
Where it sits in the network
#How this signal connects to neighbouring technologies, skills and background knowledge. Follow any link to keep exploring.
Further thought
Follow-up questions
#- 你的输入 token 里有多少按缓存价计费？降价对你的账单实际意味着多少？
- 你在用的推理档位，平均会想多久、会不会撞上输出上限？
- 顶档没降价而次顶档降一半，你有多少任务其实用次顶档就够？
Publication record
Featured in
#The briefings that carried this signal, newest first; the last row is the weekly review of its week. The tag is the section it ran in.
技术简报 - 2026-09-24
Immediate attentionWeekly review · Sep 21 – Sep 27
Worth tracking
Sources