Abliterated GLM-5.3 model launches for offensive cyber, ranks #3 on Terminal-Bench 4.0
petrusenko_max · x · 2026-09-03
abliterationai released abliterated-model-large-v2, an abliterated GLM-5.3 that ranks #3 on Terminal-Bench 4.0 behind Opus 5 and Fable. It doubles the cyber exploitation capability of v5.2 and targets offensive cyber, red teaming, and agent testing tasks other models refuse. It offers US hosting, FP8 precision, a 1M token context window, and zero prompt retention.
Related event: Abliterated GLM-5.3 Model Release Sparks Cybersecurity Policy Concerns(9 posts)→
More from Models
- Running Ling 3.1 Flash (560B MoE, 25B Active) Across 4 DGX Sparks via llama.cpp RPC — BerryBlush_Rogers691 · 2026-10-18
- Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18
- Polymarket opens betting market on who releases the next Gemini Argon model — Polymarket · 2026-10-18
- Grok matter-of-factly lists which groups user was racist against, exempting Africans and Latin Americans — teortaxesTex · 2026-10-18
- A quick test for benchmaxxed demos: swap in Mirror's Edge and see if it generalizes — teortaxesTex · 2026-10-18
- Did OpenAI distill DeepSeek's gains? Benchmark shows OpenAI's Luna scoring just 38 — teortaxesTex · 2026-10-18