WHAT NEW IN X10.29 THU

AI collections from X · 22 collections · 60,636 posts

Back to No.021 MiniMax M3.1 Flash in practice: tests and comparisons with DeepSeek V4.1 FlashUpdated 5 hours ago
minimax-m3.1IMAGE

MiniMax M3.1 Flash burns 89M tokens on a single prompt with 98% cache hit

The author says MiniMax M3.1 Flash used 89 million tokens on one task, with 570K output tokens and a 98% cache hit rate, and teases what it built.

Sahil Panhotra | Indie Builder
@SahilPanhotraArchived 7 days agoOriginally posted 2026.10.21
MiniMax M3.1 Flash burns 89M tokens on a single prompt with 98% cache hit
Sahil Panhotra | Indie Builder @SahilPanhotraOriginally posted 2026.10.21
Open original on X

I THOUGHT DEEPSEEK USING 82 MILLION TOKENS WAS INSANE 😭 MiniMax M3.1 Flash just burned through 89 MILLION TOKENS on a single prompt. 98% cache hit, 570K output tokens and all of this for ONE task 💀 wait till you see what it built

This post quotes another post: x.com/sahilpanhotra/status/2112072853894119632

x.com/sahilpanhotra/status/2112805836393480650Archived 2026.10.22 12:08
Open original on X