GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
chipherndon
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
chipherndon
Model Tree
infly/Infinity-Parser2-Flash
Input Modalities
Output Modalities
Supported Functionality
Server-side copy of infly/Infinity-Parser2-Flash at
cfddc4106b0abc4706f05575a164f7dd35d09c46.
The only change is in tokenizer_config.json:
markdown
tokenizer_class: TokenizersBackend -> Qwen2Tokenizer
tokenizer.json and the weights are unchanged. Use this repo when a serving
runtime cannot rewrite the snapshot (vLLM 0.17.x / Transformers 4.x cannot
import TokenizersBackend).
Do not swap in Qwen/Qwen3.5-2B as the model. Point --model at this repo.