Atria Dawn Preview — Hardware Requirements & GPU Compatibility
ChatFunctions
Atria Dawn Preview is a preview-stage agentic model from the Shanghai Artificial Intelligence Laboratory, built on a 744-billion-parameter Mixture-of-Experts GLM-5.2 foundation. It is an instruction-tuned model designed for research and engineering workflows that require continuous environment understanding, tool use, and multi-step task completion, spanning research-style discovery, building software and applications, producing structured deliverables like reports, and authorized cybersecurity analysis. It supports a 256K token context window and is released under the MIT license. At roughly 744 billion total parameters, local deployment at 4-bit quantization needs several hundred gigabytes of memory, putting it solidly in multi-GPU server territory; most users will reach it through a hosted endpoint rather than local hardware.
Atria Dawn Preview runs on a GPU cluster
BF16 needs ~1.5 TB · about 3 nodes of 8× H100 SXM 80 GB
MIT
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Atria Dawn Preview
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve internlm/Atria-Dawn-Preview --max-model-len 16384 --tensor-parallel-size 8 --pipeline-parallel-size <nodes>Size the nodes from the table above
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Rent 3× 8-GPU H100 SXM nodes · BF16
- Estimated hourly price
- ~$64.56/hr
- est., whole cluster
- GPUs
- 24 GPUs
- 1.9 TB of VRAM
$5+ bonus for new users
How Much VRAM Does Atria Dawn Preview Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | Runs on | |
|---|---|---|---|---|
| 1510.9 GB | 1509.9 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Runs on 3 nodes of 8× H100 SXM |
Selected BF16
Brain floating point 161508.4GBat 16K context
- Weights
- 1506.7 GB
- KV cache
- 1.5 GB
- Overhead
- 0.3 GB
Server-class: ~1.5 TB at 16K context · about 3 nodes of 8× H100 SXM 80 GB
Which Hardware Can Run Atria Dawn Preview?
BF16 · 1510.9 GBAtria Dawn Preview (BF16) requires 1510.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 1965+ GB is recommended. Using the full 1049K context window can add up to 90.3 GB, bringing total usage to 1601.2 GB. No consumer GPU has enough memory.
Needs server-class hardware
Atria Dawn Preview runs on a GPU cluster · BF16 needs ~1.5 TB
| Hardware | Quant | Memory | Context | Hardware page | |
|---|---|---|---|---|---|
| GPU cluster · 3 nodes of 8× H100 SXM 80 GB | BF16 | 1508.4 GB | 32K ctx | H100 SXM → | BF16 · 1508.4 GB · 32K ctx |
By VRAM tier · typical GPUs and Macs at BF16
Consumer GPUs (up to 32 GB) and Macs up to 512 GB: too small for this model (needs ≥ 1.5 TB)8 GB tier →12 GB tier →16 GB tier →24 GB tier →Macs →
Similar models
2| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017DeepSeek | 134.4 | 458.4 GB | —needs > 24 GB | same size |
| GLM 5.3Z.ai | 130.1 | 453.8 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- InternLM
- Parameters
- 753.3B
- Active parameters
- 41.8B per token
- Architecture
- GlmMoeDsaForCausalLM
- Context Length
- 1,048,576 tokens
- Vocabulary Size
- 154,880
- Release Date
- 2026-09-11
- License
- MIT
Frequently Asked Questions
How much VRAM does Atria Dawn Preview need?
Atria Dawn Preview requires 1510.9 GB of VRAM at BF16. Full 1049K context adds up to 90.3 GB (1601.2 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 753.3B × 16 bits ÷ 8 = 1506.7 GB
KV Cache + Overhead ≈ 4.2 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 94.5 GB (at full 1049K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Atria Dawn Preview?
No — Atria Dawn Preview requires at least 1510.9 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
Can I run Atria Dawn Preview on a Mac?
Atria Dawn Preview requires at least 1510.9 GB at BF16. None of the 146 Macs we list has that much usable memory (the largest has 508.5 GB after macOS overhead).
Can I run Atria Dawn Preview locally?
Yes — Atria Dawn Preview can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
What's the download size of Atria Dawn Preview?
At BF16, the download is about 1506.66 GB.
Which GPUs can run Atria Dawn Preview?
No single consumer GPU has enough VRAM to run Atria Dawn Preview at BF16 (1510.9 GB). Multi-GPU or professional hardware is required.
Which devices can run Atria Dawn Preview?
Atria Dawn Preview requires at least 1510.9 GB at BF16, which exceeds the usable memory of every unified-memory device we list (the largest has 508.5 GB).