GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
vishvananda
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
vishvananda
Model Tree
Qwen/Qwen3-4B-Instruct-2507
Input Modalities
Output Modalities
Supported Functionality
Research QLoRA adapter; epoch 1.2500. This is a checkpoint, not a claim that training or evaluation is complete. No final-test accuracy is claimed.
Dataset: vishvananda/mtg-oracle-design-descriptions-v2@7c1bbca04b4259c15c397c5c6a00af5e143ef21e. Exact base revision: cdbee75f17c01a7cc42f958dc650907174af0554. This snapshot is tagged epoch2-quarter-train-step-00024718. Pin that tag/commit when downloading. Download the base separately; load with Transformers/PEFT, apply system-prompt.txt, and disable thinking. See checkpoint.json for the recipe and data/code identities.
Base attribution: Qwen team, Apache 2.0. The dataset includes third-party Magic: The Gathering card content; retain its DATA_LICENSE.md and Wizards of the Coast/Scryfall/MTGJSON attribution. No artwork, optimizer state, credentials, or worker logs are included.