GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
GammaAGI
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
GammaAGI
Model Tree
Base
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
Adapter
this model
Input Modalities
Text
Output Modalities
Text
Supported Functionality
Dedicated Endpoints
Train a bounded, traceable EvoStream adapter on top of the accepted Nano scaffold using verified donor, code, reasoning, safety, and Omni contracts.
OK2.597652.559422[1.429411, 2.185695, 1.758911, 3.044919, 1.152849, 1.433145, 0.960366, 0.732565, 1.325378, 2.308462, 2.535948, 1.287751, 1.272575, 1.609289, 4.81366, 1.082453, 4.178133, 4.02433, 4.390209, 4.39222, 3.61718, 2.580169, 3.106896, 3.954231, 3.766961, 1.590856, 3.03544, 2.990237, 3.60154, 3.90493, 4.112782, 0.776599, 1.386399, 2.127077, 1.716018, 3.047485]1104842240.00348658/workspace/gamma_30b_alpha_lora/input/gamma30_v2_4_1_qwen36_compact_repair_pack_20260615.jsonl32Run strict v2.1 evaluation, then merge/unload or consolidate only if accepted.