GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
WaveCut
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
WaveCut
Model Tree
Base
Qwen/Qwen3.6-35B-A3B
Fine-tuned
this model
Input Modalities
Text
Image
Video
Output Modalities
Text
Supported Functionality
Dedicated Endpoints
Qwen/Qwen3.6-35B-A3B84a3030716a0059589e9d10e2ea049e32b76cfa65.12.1160code+math+agent+ru0.20,0.10,0.35,0.353072, 512, 42grouping=ream, saliency=reap, merging=logits+weights, group_size=32The main language MoE layers were REAM-merged to 160 experts. Qwen3.6 MTP tensors are packed differently from the upstream REAM MTP helper assumptions, so packed MTP tensors are reduced with a saved REAM group map rather than upstream --mtp_safe_tensors.
Structural audit:
160 experts160 expertsmodel.safetensors.index.json includes mtp.* keysqwen3_next_mtp drafter load must pass before the checkpoint is acceptedThe recipe favors:
zake7749/Qwen-3.6-plus-agent-tool-calling-trajectoryZeroAgency/ru-big-russian-datasetAgent calibration had 2477 unique accepted samples and was deterministically repeated to fill the 3072 batch. RU calibration had 3072 unique accepted samples.
Artifacts in this repo include:
stats/structural_audit.jsonstats/smoke_outputs.jsonstats/agent_calibration_stats.jsonstats/ru_calibration_stats.jsonlogs/main_merge.logmanifests/hf_final_audit.jsonSmoke prompts covered Russian Cyrillic chat, coding-agent debugging, and XML-style tool calling.