--- license: other license_name: edge0-community-model-1.0 license_link: LICENSE library_name: edge0 tags: - moe - edge-inference - latent-moe - 4-bit - agent - coding - cuda base_model: - Qwen/Qwen3.8-Flash-Next pipeline_tag: text-generation ---

Edge0

Edge0-145B · NVFP4

Think big. Run here. A 145B agent. On one consumer GPU. Sparse. By design. Latent MoE. Trained natively at 4 bits.

Quality · Speed · Architecture · Quick start

A 145B agent, running on one consumer GPU.

98.3 tok/s agent decode · 3165.7 tok/s 16K prefill · 86.5 terminal-bench 2.1.

Throughput measured on RTX 5090 D 24 GB · NVFP4.

A latent sparse MoE: 48 layers, 512 experts, 10 active per token, 256K context, trained natively at 4 bits.

This repository: Edge0-145B · NVFP4

ⓘ With residency enabled, 145B NVFP4 used 23.25 GiB VRAM + 38.27 GiB host RAM on RTX 5090 D 24 GB.

Model overview

NVFP4 · RTX 5090 D 24 GB. Prefill throughput uses a 16K-token input; weight sizes are listed separately for each format.

Quality

Higher is better. Results are listed by model.

Speed

Throughput is shown in tok/s, grouped by hardware and build. The summary above uses NVFP4 on RTX 5090 D 24 GB.

Full speed table · all reported machines and builds

Highlighted rows show the NVFP4 build in this repository.

Machine	Build	4K prefill	16K prefill	Chat decode	Agent decode
RTX 5090 D 24 GB	Q3_K	2169.9	3554.9	99.7	121.2
RTX 5090 D 24 GB	NVFP4	1839.6	3165.7	79.6	98.3
RTX 4090 24 GB	Q3_K	2042.0	3179.4	69.9	91.7
RTX Spark N1X 60 GiB	Q3_K	674.0	770.3	37.7	36.7
RTX Spark N1X 60 GiB	NVFP4	821.2	922.4	40.1	46.0
Apple M2 Ultra (Metal)	Q3_K	755.6	800.3	60.9	57.0
16 GiB VRAM + 20 GiB RAM	Q3_K	249	335	36.8	44.9

Measurement conditions · constrained budget and memory residency

16 GiB VRAM + 20 GiB RAM rows are the RTX 5090 D 24 GB again, constrained by cgroup to a 16 GiB VRAM and 20 GiB system-RAM budget — not a different card. With residency on, the same card holds Q3_K (40.81 GiB of weights) as 23.14 GiB VRAM + 31.28 GiB RAM, and NVFP4 (47.79 GiB) as 23.25 GiB VRAM + 38.27 GiB RAM — disjoint expert sets, nothing counted twice.

These residency readings use expert residency enabled, --no-prefill-borrow, conv cache = 0 and kv-resident = 0. Host RAM is measured as RssAnon + RssShmem. These are single-run readings, not minimum system requirements. See Usage & troubleshooting for memory-saving options.

In action

Recorded with Edge0-145BOpen video ↗

Architecture

Architecture details · diagram, configuration and expert routing

	Edge0-145B
Total parameters	145B · 10.12 B active per token
Layers	48 · hybrid attention, full every 4 layers + sparse indexer
Experts	512 per layer, 10 routed + 1 shared active per token, SwiGLU width 640
Expert pool	84.6 B parameters at latent width ℓ = 1792
PLE	51.2 B n-gram embedding table, fp8, memory-mapped
MTP head	1.26 B draft model, 512 experts, for speculative decode
Context	256K · KV int8 with 32K resident
Builds	Q3_K 40.81 GiB · NVFP4 47.79 GiB

Left: the stack the model actually runs — token embedding, 48 hybrid-attention layers, final norm, lm_head. Right: one layer opened up. Attention reads and writes the full 2560 width. D then projects the token down to ℓ per routing group, the router picks 10 of 512 experts, each expert runs ℓ → 640 → ℓ, a shared expert runs alongside, and U projects the combined result back to 2560.

Everything outside the expert pool stays at 2560: D and U are shared, one copy per layer. Only the 512 experts work at ℓ. That is why an expert is a fixed 1,630,720 B record — the engine keeps a few thousand of them on the card and streams the rest from host memory, at a cost that tracks how many experts are resident and not how big the model is.

Quick start

Each model folder (Q3_K or NVFP4, Edge0-145B or Edge0-120B) runs on Edge0 through edge0-infer, from the command line or from the Edge0 Chat desktop app. Keep the folder as it is: edge0-infer identifies the model by model_name in its model.json.

Platform & documentation notes

The build instructions target NVIDIA GPUs with CUDA 13 on Linux or Windows (x64 or ARM64). The Apple M2 Ultra measurements use Metal; they do not establish macOS release availability.

Build & command-line instructions

1. Build. Clone the repository, then build two parts:

git clone https://github.com/Edge0-AI/Edge0.git edge0

edge0-strata, the engine: follow "Build (Linux, NVIDIA)" or "Build (Windows, NVIDIA)" in edge0-strata/EDGE0.md.
edge0-infer: follow "Build" (Linux) or "Windows" in edge0-infer/docs/getting-started.md.

2. Run from the command line. Pass one model folder, such as Edge0-145B-NVFP4, not the folder that holds several of them. On Windows the program is build\windows\bin\edge0-infer.exe.

build/cpu/bin/edge0-infer serve --model /path/to/Edge0-145B-NVFP4

It prints Listening: http://127.0.0.1:8000 model=<name> once the weights are loaded, where <name> is edge0_145b, and then serves OpenAI- and Anthropic-compatible endpoints:

curl localhost:8000/v1/chat/completions -H 'content-type: application/json' -d '{"model":"edge0_145b","messages":[{"role":"user","content":"hi"}],"max_tokens":64}'

edge0-infer chat --model /path/to/Edge0-145B-NVFP4 opens an interactive session instead. More options: edge0-infer/docs/edge0-145b.md.

Desktop app · Windows

⬇️ Download: GitHub Releases

Open Edge0 Chat on Windows and choose a model folder, such as Edge0-145B-NVFP4, to run the model locally. To build the app from source, follow the Edge0 Chat guide.

Usage & troubleshooting

145B / 120B · Memory options

Low RAM · SSD offload

SSD mode: --mmap-experts — read uncached experts from SSD on demand.
Combined mode: --resident-budget-gib 20 + static --expert-profile — SSD loading with a host expert-cache budget and disjoint GPU/RAM caches. Includes --mmap-experts automatically.

RAM budget · what 20 GiB covers

Expert cache: --resident-budget-gib 20 caps host-RAM expert caching at 20 GiB, lowered further to fit available memory.
Total RAM: expert cache + other runtime allocations. This flag does not set a total-process RAM limit.
Benchmark limit: the 20 GiB total RAM test budget is a separate measurement condition.

GPU / RAM · avoid duplicate experts

Enable: --resident-experts or --resident-budget-gib N, with a static --expert-profile.
Effect: experts cached in VRAM need no duplicate RAM copy. Compatible with SSD loading.
Default: off, prioritizing speed.
Automatically enabled: --mmap-experts, --resident-cpu-experts, --resident-pin.

Acknowledgements

Credits & license

We thank the teams behind Qwen3.8-Flash-Next and Strata for sharing their work with the community. Edge0 extends the Strata engine with latent MoE support.

Model license. Edge0-specific contributions are provided under the Edge0 Community Model License 1.0, only to the extent that Edge0 owns or is authorized to license them. Qwen materials remain subject to the Qwen Community License 1.0. Both sets of applicable terms must be followed; see NOTICE and the release scope.

The Edge0 license permits personal use, research, modification, noncommercial sharing, internal use, and commercial use of generated outputs, subject to applicable upstream conditions. Selling or commercially bundling the model, operating a commercial model service, or offering an independent commercial coding or office assistant using it requires a separate written Edge0 license. See LICENSE for the precise scope and exceptions.

Commercial licensing: info@edge0.ai. Edge0 permission does not waive Qwen's separate authorization or attribution requirements. Framework code retains its Apache-2.0 license; Strata retains its MIT license. Existing valid grants are not revoked.