GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
promotion
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
promotion
Model Tree
meta-llama/Llama-3.1-8B-Instruct
Input Modalities
Output Modalities
Supported Functionality
Research checkpoint for the RONPO AAAI revision experiments.
meta-llama/Llama-3.1-8B-Instruct/NHNHOME/AIPR/sjkim/MNPO_rev_20260710/results/p8_stage4_fresh_default_test_20260718/stage4/ronpo_os_stage4/train/fullnot foundMeasured P8 Stage-4 result on the locked 1,000-prompt PKU-SafeRLHF default-test panel: RONPO OS passed the corrected stability gate and exceeded IPO on Worst by +0.0445 (paired bootstrap 95% CI [+0.0217,+0.0670]); its Avg exceeded SimPO by +0.0248 (95% CI [+0.0082,+0.0416]). The panel is prompt-disjoint from earlier listed panels; no spent sealed split was touched.
Intended use: reproducibility and evaluation for the RONPO paper. This checkpoint is not intended as a production assistant.