GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
DavidAU
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
DavidAU
Model Tree
Input Modalities
Output Modalities
Supported Functionality
This model was used as a part of:
This version exceeds "Qwen3.6-27B-F451-AND-TRI-Polar-Ultra-Pro-Writer-Uncensored-Heretic" in ALL BENCHMARKS, including scoring OVER 700 in ARC-C for both 8 bit and 4 bit.
(see the benchmarks on the "711" repo.)
An Unsloth fine tune on local hardware to specifically raise the model's core intelligence.
markdown
Benchmarks:arc arc/e boolq hswag obkqa piqa winoQwen3.6-27B-F451-AND-TRI-Polar-Ultra-Pro-Writer-Uncensored-Heretic [instruct]mxfp8 0.665,0.830,0.909,...mxfp4 0.654,0.830,0.905,...Qwen3.6-27B-Instruct: [base, non heretic]qx86-hi 0.637,0.798,0.911,0.775,0.442,0.807,0.737Qwen3.5-27B-Instruct: [base, non heretic]mxfp8 0.557,0.711,0.868,0.533,0.452,0.706,0.695
Models are tested in "Instruct" mode because this generally works better with the testing harness.
Testing via "thinking" mode also shows the metrics (and changes) but not the true extent.