Platform · Hardware and models
Verified silicon, targets in bring up, and every recipe we ship.
The same Rust and CUDA source runs on NVIDIA and AMD without a second kernel tree. This page names what is verified, what is in bring up, and every model recipe, generated from the repository at build time.
Who this is forEngineers checking whether their hardware and their models are covered before a pilot.
Verified hardware
GB10 · Blackwell SM121
VerifiedNVIDIA DGX Spark
One multi model binary serves a full matrix of hand tuned targets on a single GB10. NVFP4 and FP8, MTP speculative decoding, EP=2 across two Sparks. Every target passes the serve matrix before an image is cut.
Deployment guide ↗gfx1151 · RDNA 3.5
MLPerf submittedAMD Strix Halo
One codebase, both vendors. The CUDA kernels compile straight for gfx1151 through SCALE by Spectral Compute. AMD provided the Strix Halo desktop the MLPerf Inference v6.1 numbers were submitted on.
About SCALE ↗Hopper · SM90
Bring upNVIDIA H100 and H200
Hopper owned decode and prefill kernels with published receipts in the changelog, bit identical to the reference on the production shapes. Not yet a verified target, and this page will say so until it is.
Blackwell · SM100
CampaignNVIDIA B200 and GB200
Datacenter Blackwell is the next objective after Hopper, built around the same principle the hardware demonstrates, inference economics improve when the software is designed for the accelerator.
gfx1201 · RDNA 4
In reviewAMD Radeon AI PRO R9700
A SCALE target serving Qwen3.8-27B in an open pull request. The compatibility list grows in the open.
Battlemage
In talksIntel Arc Pro B70
Active conversations with Intel. Nothing is signed and this card will say so until it is.
Models
Every model here has a recipe.
Pick a vendor, then a family. Every card maps to one recipe in atlas-recipes, so the site cannot list a model it does not ship.
Qwen3-Coder-Next
atlasctl run qwen3-coder-next-fp8Qwen3-Next
atlasctl run qwen3-next-80b-a3b-nvfp4Qwen3-VL
atlasctl run qwen3-vl-30b-a3b-nvfp4Qwen3.5
atlasctl run qwen3.5-0.8b-bf16-atlasatlasctl run qwen3.5-122b-a10b-nvfp4-ep2 --rank 0 --world-size 2 --master-addr <spark-1>atlasctl run qwen3.5-122b-a10b-nvfp4-singleatlasctl run qwen3.5-27b-dense-nvfp4atlasctl run qwen3.5-35b-a3b-nvfp4Qwen3.6
atlasctl run qwen3.6-27b-fp8atlasctl run qwen3.6-27b-fp8-mtpatlasctl run qwen3.6-27b-nvfp4atlasctl run qwen3.6-27b-nvfp4-prefill-recordatlasctl run qwen3.6-27b-nvfp4-unslothatlasctl run qwen3.6-35b-a3b-fp8-bf16headatlasctl run qwen3.6-35b-a3b-fp8-mtpatlasctl run qwen3.6-35b-a3b-fp8-nvfp4headatlasctl run qwen3.6-35b-a3b-nvfp4Qwen3.8
atlasctl run qwen3.8-27b-nvfp4-dflash2atlasctl run qwen3.8-27b-nvfp4-latencyatlasctl run qwen3.8-27b-nvfp4-throughputatlasctl run qwen3.8-27b-nvfp4-unslothatlasctl run qwen3.8-27b-nvfp4-unsloth-bfclatlasctl run qwen3.8-flash-next-nvfp4Questions
The questions we actually get asked.
Short answers. Each one is backed by something on this site or in the repository.
What hardware does it run on?
NVIDIA DGX Spark (GB10) is verified today, and AMD Strix Halo (gfx1151) runs the same CUDA source compiled through SCALE, with both submitted to MLPerf Inference v6.1. Hopper and Blackwell datacenter targets are in active bring up with receipts in the changelog. Expert parallelism across two nodes ships as recipes and a three node topology is being wired up.
Which models can I run?
Every model on this site maps to a recipe in the atlas-recipes repository, which is the single source of truth, so the site cannot list a model without one. Qwen leads with the most recipes, alongside Gemma, Nemotron, Mistral, MiniMax and DeepSeek. Bring your own weights and we scope the bring up.
Ask the rest in a working session, or read the deployment guide ↗.
Next step
See it against your own workload.
A side by side ladder on your hardware in week one. Your models, your criteria, your receipt.