The short version
OpenAI's cost-efficient GPT-5.6 model running at maximum reasoning effort.
From the HackerLinks archive
GPT-5.6 Luna Max
OpenAI's cost-efficient GPT-5.6 model running at maximum reasoning effort.
At a glance:
First seen:2026-09-17
Last seen:2026-09-17
Times seen:1
Website:developers.openai.com
Why it caught our attention
A Rust developer was very happy using it for asynchronous subagent reviews.
Where it surfaced on Hacker News
2026-09-17
“I use Luna Max for asynchronous subagent reviews and am very happy with its work, but it’s slow af.”
rapind · recommendation · first hand use
Direct HN comment
Xiaomi Mimo 2.6 live post-training dashboard
Also surfaced in this discussion
Xiaomi MiMo-V2.5
Xiaomi Mimo 2.6 live post-training dashboard
2026-09-17