Thinking Machines' Inkling 2 Mini matches Claude Opus 5.5 at 74.58% on CyberBench
LiTianleli · x · 2026-11-21
A CyberBench v1.1 (VALS, "finding and patching security bugs") leaderboard screenshot shows Thinking Machines' small open model Inkling 2 Mini scoring 74.58% — tied with Claude Opus 5.5 (CVP) at ranks 9-10 — at only $0.93 per test versus Opus's $4.17.
The top of the leaderboard: Wonder at 80.36%, GPT-6 Sol at 77.98%, Gemini 4 Argon at 77.86%. MiMo V2.6 Flash hits 75.36% for just $0.05 per test. Multiple users report Inkling-2-mini outputs beat Qwen 27B quality at much higher speed, calling it a rare strong result for a US small open model.
More from Models
- LeCun: Training LLMs on human knowledge yields memory, not intelligence — haider1 · 2026-11-22
- Claude veteran shocked by Gemini's leap, weighs switching to Google AI Ultra — ScreenPlayLife · 2026-11-22
- Gemini Neon builds an interactive USL lock-contention UI exposing a testing blind spot — tianyin_xu · 2026-11-22
- New spatial planning benchmark: GPT-6 and Claude Opus 5.5 fail parking route tasks — songhan_mit · 2026-11-22
- Microsoft's new models reportedly up to 80x cheaper and 50x faster — ChrisUniverse · 2026-11-22
- Inkling 2 Mini uncensored: refusals 100%→0% with a 40MB LoRA, 98% MMLU kept — Top-Eye-8104 · 2026-11-22