Abliterated-model-large removes refusals from GLM-5.2 and still scores well on coding tests
Effective_Attempt_72 · reddit · 2026-07-26
The team says it shipped abliterated-model-large, a GLM-5.2 variant with refusal directions removed and then fine-tuned for long adversarial and agent-style tasks.
Reported evals include 84.2% on CyberGym, 86.2% AgentHarm compliance with zero refusals on the published set, 97.5% AgentDojo utility, 81.2% SWE-bench Verified, and 80.1% Terminal-Bench 2.1.
It is exposed as an OpenAI/Anthropic-compatible API with zero data retention by default. The authors say the model itself has no built-in policy and that users control the rules.
Related event: Abliterated GLM-5.2 Variant Released for Red Teaming(2 posts)→
More from Models
- Harness bug masked Gemini 4 Argon: rerun with Antigravity beats Astra, nears Opus 5.5 — vitaliychiley · 2026-10-27
- GPT 6.1 Terra vs Claude Sonnet 5.5: speed and value vs raw capability — teortaxesTex · 2026-10-27
- Harmonic Teases: AI Taught to Prove Math, 'Something Harder' Coming Soon — satnam6502 · 2026-10-27
- OpenAI quietly swaps Plus/Pro default to cheaper GPT-6.1 Terra and cuts Codex allowance — kimmonismus · 2026-10-27
- 'Never been this uninterested after an OpenAI model release' — theteknosaur · 2026-10-27
- Dev questions reports of Chinese models writing insecure code, citing thin methods and conflicts — dbreunig · 2026-10-27