GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
DJLougen
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
DJLougen
Model Tree
GestaltLabs/Ornstein-3.5-9B-V2
Input Modalities
Output Modalities
Supported Functionality
| Qwen3.5-9B-Base | Ornstein V1.5 | Ornstein V2 | V2-Coder (exp.) | |
|---|---|---|---|---|
| Overall | 0.725 | 0.850 | 0.825 | 0.785 |
| Reasoning | 0.68 | 0.90 | 1.00 | 1.00 |
| GPQA | 0.36 | 0.80 | 1.00 | 1.00 |
| Coding | 0.77 | 0.80 | 0.65 | 0.57 |
(V1.5 column = published reference; V2/V2-Coder = fresh same-seed runs. Coding = livecodebench + hlce.)
Retains V2's vision tower + MTP head (full multimodal weights).
python
from transformers import AutoModelForCausalLM, AutoTokenizerimport torchmodel_id = "DJLougen/Ornstein-3.5-9B-V2-Coder-experimental"tok = AutoTokenizer.from_pretrained(model_id)model = AutoModelForCausalLM.from_pretrained(model_id, dtype=torch.bfloat16, device_map="auto")
I'm a PhD student in visual neuroscience at the University of Toronto who also happens to spend way too much time fine-tuning, merging, and quantizing open-weight models on rented H100s/B200s and a local DGX Spark. All training compute is self-funded. If my uploads have been useful, consider buying a PhD student a coffee.
Apache 2.0 — inherited from the Qwen 3.5 9B base release.