edge0

The world’s first on-device AI that runs on any hardware

Runs on low-power hardware | Cloud-level intelligence | Works offline | On-device data processing | Zero token cost

Explore our models

Unlimited access to advanced AI

Inference runs on the hardware you already ship, resulting in no per-use charges and no costs that scale with usage.

Designed for low-power, low-memory devices

Run a 35-billion-parameter language model with a small active memory footprint by streaming expert weights from SSD. Across the Edge0 family, language and voice models support products from phones and laptops to earbuds, glasses, and toys.

Private by design

Run inference on-device to keep sensitive data on the hardware, without sending it to a cloud model for processing.

How far down the device ladder
can on-device AI actually reach?

Most stay at the research stage

A few reach high-compute hardware

A handful stop at mid-compute hardware

Edge0All Four Tiers

Default state: no stage gradient background shown.

Edge0 on-device model matrix

✦ Cloud-level intelligence · ✦ Works offline · ✦ Zero token cost · ✦ Data stays on-device

Edge0 pioneered 10B+ and 100B+ on-device models, breaking the memory wall and taking cloud-class AI from research to production.Compatible with 95% of mainstream chips, they run on phones, PCs, wearables, IoT and robots with low memory use and power draw.With over 110,000 customers served, we're working toward one vision: Intelligence at Zero Distance.

200+

On-device models

5M+

Devices in production

15M+

Open-source downloads in 3 months

On-device foundation models

Edge0

Vibe codingAgentic tasks

Laptop-ready · Vibe coding

145B / 120B

World’s first 100B+ parameter on-device model

Phone-ready · Agentic tasks

35B / 8B

#1 on a Hugging Face leaderboard
World’s first 10B+ on-device model

Data from company papers and internal tests. Hugging Face rankings reflect results at release.

Parameter count is not the same thing as RAM

See the models

Edge0-35B · sparse MoE

1.06 GB

LFM2-2.6B

5.4 GB

MiniCPM-V 4.0

9.0 GB

Gemma 4 E2B

11.4 GB

MiniCPM4-8B

16.4 GB

Gemma 4 E4B

17.9 GB

Qwen3.8-27B

56.0 GB

Muse Glimmer 30B

56.0 GB

Qwen3.5-35B-A3B

70.0 GB

DeepSeek-R1-Distill-Llama-70B

140.0 GB

A family of models.
One engineering discipline.

All models and specs

AI that consumer hardware can actually run.

Explore Edge0 products, evaluate them on your hardware, or talk to our team about integration.

SSD / FLASHFull ModelRAMCPU / NPUComputeShared CoreSLOT A · Current LayerSLOT B · Next LayerPLANNEROn-Demand LoadPrefetch Next LayerRelease Computed Layer

Subscribe for the latest news on Edge0 products