GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
CloudGoat
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
CloudGoat
Model Tree
Qwen/Qwen3.5-4B
allenai/tmax-4b
InternScience/Agents-A1-4B
Input Modalities
Output Modalities
Supported Functionality
This model was merged using the DARE TIES merge method using Qwen/Qwen3.5-4B as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
yaml
merge_method: dare_tiesbase_model: Qwen/Qwen3.5-4Bmodels:- model: InternScience/Agents-A1-4Bparameters:weight: 1.0density: 1.0- model: allenai/tmax-4bparameters:weight: 0.7density: 0.3parameters:int8_mask: truenormalize: falsedtype: bfloat16
| Tasks | InternScience/Agents-A1-4B | CloudGoat/Mephisto-4B-0725 |
|---|---|---|
| JCommonsenseQA (acc) | 86.00% (±2.01%) | 84.67% (±2.08%) |
| JNLI (acc) | 76.33% (±2.46%) | 77.00% (±2.43%) |
| JSQuAD (Exact Match) | 65.33% (±2.75%) | 64.67% (±2.76%) |
| MARC-ja (acc) | 95.00% (±1.26%) | 95.00% (±1.26%) |
| Ja-Overall (Average) | 80.67% (±0.00%) | 80.34% (±0.00%) |
| GAIA Level 1 | 12/53 |