As of: 2026-10-10 14:45 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code). Human editor: Adam Bicz.
Canonical page: https://postcutoff.com/p/x-envoldev-2107973447230091623/

# “My famous night train benchmark is in!!! 🚂🌙🔥 Haiku 5.5 vs.”

Tony (@EnvolDev), X, 2026-10-07.
Source: https://x.com/EnvolDev/status/2107973447230091623

## Archived text

> My famous night train benchmark is in!!! 🚂🌙🔥
> Haiku 5.5 vs. GPT-6 Luna ⚔
> 
> Honestly, Haiku 5.5 completely failed this test. There’s no comparison. Luna is miles ahead!
> 
> ⚡ GPT-6 Luna: 54.6 seconds for under $0.01.
> 🐌 Haiku 5.5: 5 minutes 12 seconds for $0.03.
> 
> Luna was over 5× faster and cost less than a third as much—with a much better result! 🤯

Archived 2026-10-10 via fxtwitter (unofficial).
Counts: 32,966 views, 350 likes, 13 reposts, 52 replies (at fetch time)
Media: https://video.twimg.com/amplify_video/2107973109445959680/vid/avc1/1080x1080/RxExMNkEULBH_QrL.mp4?tag=29

## Cited in

- Video: [Night-train counterexample](https://postcutoff.com/v/envoldev-haiku-5-5-vs-gpt-6-luna-night-train-x/)