Inkling-2 Mini adapts its thinking depth to task size, runs locally at 91 tok/s on RTX 3090
LiTianleli · x · 2026-11-22
Hands-on with Thinking Machines' Inkling-2 Mini highlights its adaptive thinking budget: small tasks finish thinking in under 10k tokens, bigger prompts get more reasoning without looping — unlike many models that apply uniform effort regardless of task size. It runs fully locally (not via API): the Unsloth UD-Q4KXL quant on an RTX 3090 via superalesha's Maya fork at 91 tok/s, with the fork merging into mainline Maya soon. Screenshot shows 0 failed runs, 1.7k tok/s prefill, one 24-minute answer with 67,432 thinking tokens.
More from Models
- She churned through a week's Claude usage researching the Theranos movie — DavidWells · 2026-11-22
- Custom Rust engine runs 524k context on one RX 7900 XTX, 5x faster than llama.cpp — W61k3r · 2026-11-22
- Fine-tune gotcha: HF export silently drops Qwen3's MTP head, slowing decode — AIFlow_ML · 2026-11-22
- Designer KOL: Google's model pace and quality suggest it will soon 'mog them all' — AIandDesign · 2026-11-22
- Models play a literary 'listening at doors' game — Opus 4.5 gets recognized by its timbre — repligate · 2026-11-22
- Asking Claude to 'understand the entire mathematical platonic realm, make no mistakes' — burny_tech · 2026-11-22