GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
tolegend
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
tolegend
Model Tree
Input Modalities
Output Modalities
Supported Functionality
This model was merged using the Task Arithmetic merge method using /root/bittensor/sn120/models/255 as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
yaml
# Task Arithmetic merge: 255 is the strong base; mine's SWE advantage is# extracted as a task vector and added on top of 255.# This preserves 255's strength on MEMORY/NAVWORLD/TERMINAL while# injecting mine's SWE capability.## How to tune:# scaling_coefficient: 0.1–0.4 (low because mine only adds SWE, not overall quality)# Start at 0.3, raise if SWE is still weak, lower if other envs regress.merge_method: task_arithmeticbase_model: /root/bittensor/sn120/models/255models:- model: /root/bittensor/sn120/models/255parameters:weight: 1.0- model: /root/bittensor/sn120/models/mineparameters:weight: 1.0parameters:scaling_coefficient: 0.3dtype: bfloat16tokenizer_source: base