dgx-spark-qwen38
GitHub

dgx-spark-qwen38 / Models / Ornith 1.5 122B

Run Ornith 1.5 122B on a DGX Spark

A 122B mixture of experts with vision, by Ornith AI. NVFP4 with its MTP head, on vLLM.

Ornith AI's 122B mixture of experts with vision and a window of a million tokens, in NVFP4 on vLLM with its MTP head. On an ASUS Ascent GX10 with 128 GB: 42.4 tok/s, greedy median, one request.

Measured

42.4tok/s

greedy median, one request

  • 90.0 tok/s in all at 8 streams on five real jobs
  • GSM8K 98.8% on 250 problems, on a second GX10
  • A window of a million tokens
Measured on
an ASUS Ascent GX10 with 128 GB, the project's reference box
How
./bench.sh
Source
docs/ornith-lane.md

Load it

From the cockpit: open the Lanes view, pick Ornith 1.5 122B and press Load (5 and a half minutes the first time, 2 to 4 after). From a terminal, in the repo folder on the box:

./prepare-lane.sh ornith-122b   # the first time only, builds the lane
./switch-model.sh ornith-122b   # switches the boot, then prints the commands to run

switch-model.sh never restarts an engine itself: it prints the exact stop and start commands. The cockpit's Load runs them for you.

On a 64 GB DGX Spark

No: it needs a 128 GB DGX Spark. A 64 GB box serves Nemotron 3.5 Bolt and Qwen3.8-27B: what runs on 64 GB.

Use it from your tools

The box serves OpenAI- and Anthropic-compatible APIs on port 30001, with an API key. Claude Code, opencode, Open WebUI and the OpenAI and Anthropic SDKs point there instead of a cloud: see Claude Code with a local model and opencode with a local model.

Install it

One command, on stock DGX OS. The same command updates it later.

curl -fsSL https://raw.githubusercontent.com/hasso5703/dgx-spark-qwen38/main/get.sh | bash

Questions about Ornith 1.5 122B on a DGX Spark

Can a DGX Spark run Ornith 1.5 122B?

Yes. Ornith AI's 122B mixture of experts with vision and a window of a million tokens, in NVFP4 on vLLM with its MTP head. The installer of dgx-spark-qwen38 sets it up, and the cockpit loads it with one button.

How fast is Ornith 1.5 122B on a DGX Spark?

42.4 tok/s, greedy median, one request, measured on an ASUS Ascent GX10 with 128 GB. 90.0 tok/s in all at 8 streams on five real jobs. GSM8K 98.8% on 250 problems, on a second GX10. A window of a million tokens.

Does Ornith 1.5 122B run on a 64 GB DGX Spark?

No: it needs a 128 GB DGX Spark.

The other lanes