GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
XCombinator
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
XCombinator
Model Tree
Qwen/Qwen2.5-1.5B-Instruct
Input Modalities
Output Modalities
Supported Functionality
Unified JSON format: a system prompt (task + output schema) + a numbered user sequence → one JSON
answer ({"reasoning": "...", "steps": [...]} for next-step/completion;
text
{"reasoning": "...", "valid": bool, "rule": "RULE_..."|null}zo_train.prompts.build_messages from the
project repo, then apply the tokenizer chat
template. See the flagship model card for a full from_pretrained snippet.
| task | this checkpoint | n-gram baseline |
|---|---|---|
| next-step (top-1) | 0.435 | 0.69 |
| sequence completion (block-acc) | 0.500 | 0.637 |
| anomaly (F1) | 0.000 | 0.89 |
Full study + all checkpoints: the project repo and submissions/XCombinator/REPORT.md.
AutoModelForCausalLM.from_pretrained.