GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
KiloHelios
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
KiloHelios
Model Tree
Qwen/Qwen3.5-9B
Input Modalities
Output Modalities
Supported Functionality
global_step_600qwen35_9b_full_grpo_song2_4plus4_len2048_nothink_save300_modelonly_newreward4d_strictbuf_val100_20260827_214901lora_rank=0)1e-60.001, low_var_klThe training reward is the equal-weight arithmetic mean of four Judge scores on the original 1–5 scale:
The run used a strict bounded reward replay buffer with capacity 1024. Judge
requests used the configured gpt-oss-20b endpoints.
This model requires a recent Transformers version with Qwen3.5 support.
python
import torchfrom transformers import pipelinepipe = pipeline("image-text-to-text",model="KiloHelios/Qwen3.5-9B-PereGRM-NewReward4D-Step600",dtype=torch.bfloat16,device_map="auto",)messages = [{"role": "user","content": [{"type": "text", "text": "How can I plan a focused week?"}],}]result = pipe(text=messages, max_new_tokens=512)print(result[0]["generated_text"][-1]["content"])
The repository includes the tokenizer, processor configuration, and the chat template saved with the checkpoint.
This is a research checkpoint optimized for a personalized-response reward. It has not been established as a general capability or safety improvement. Outputs should be independently evaluated for the intended deployment domain.