MiniMax confirms Space Bunny was early M3.1 Flash: $0.15/M input, open weights Oct 23
lmoroney · x · 2026-10-20
MiniMax confirmed that Space Bunny, the stealth model that processed over 200 trillion tokens, was an early version of M3.1 Flash. Official numbers: $0.15 per million input tokens, $0.60 per million output, 1M-token context with text, image, and video input, and open weights arriving October 23.
The author suggests builders run a small eval this week: ten tasks from your actual workload, scored identically across two or three models, so you'll know within an hour of the weights landing whether it earns a place in your stack.
More from Models
- MiniMax M3.1 Flash accused of benchmaxxing on Artificial Analysis' private Omniscience bench — teortaxesTex · 2026-10-20
- From memory capacity to transformers: outer-product stores cap out near 2×key dimension — burny_tech · 2026-10-20
- Reddit users slam Google's shrinking Gemini usage limits, suspect compute shortage — JesseBell935 · 2026-10-20
- Training a LoRA locally in 30 minutes? Veteran says you're probably lying — the_bollo · 2026-10-20
- Gemini 4 Argon finds 74% of known vulnerabilities, top recall on dfbench — andreamichi · 2026-10-20
- Frontier models already saturate persimmon, a tiny human-distribution benchmark — noahdgoodman · 2026-10-20