Pinned Tweet
A major turning point for our crew.
Frontier-level agentic coding, now running fully on-device. Edge0-145B and Edge0-120B are available with open weights.
Give them a try.
Frontier-level agentic coding, now running fully on-device. Edge0-145B and Edge0-120B are available with open weights.
Give them a try.
Introducing Edge0-145B and Edge0-120B, @Edge0_AI’s first device-native 100B-class foundation models with open weights.
Cloud-level intelligence on your laptop. Free, offline, and fast enough for real vibe coding.
- Smart: beats Claude Opus 4.6 on Terminal-Bench 2.1 (86.5% vs. ust 29G memory, 1/3 of a standard setup (100G+)
- Fast: decodes 75+ toks/s, prefills 1050+ toks/s on RTX Spark and Mac.
- Steady: handles 128K–256K tokens of context without slowing down, even on long tasks
Frontier-class agentic coding. No cloud API. No token bill. Your code and data never leave your machine.
Open weights are live on @huggingface. Run it locally or plug it into your favorite coding agent.
2
4
30
1,245
Both models and the inference engine are available now.
145B: huggingface.co/Edge0/Edge0-14…
120B: huggingface.co/Edge0/Edge0-12…
Engine: github.com/Edge0-AI/Edge0
App: github.com/Edge0-AI/Edge0…
145B: huggingface.co/Edge0/Edge0-14…
120B: huggingface.co/Edge0/Edge0-12…
Engine: github.com/Edge0-AI/Edge0
App: github.com/Edge0-AI/Edge0…
3
0
3
141
More on Edge0-145B, including the architecture and demos:
edge0.ai/blog/introduci…
edge0.ai/blog/introduci…
1
0
1
87
35B parameters on a phone. Sounds like a lot to squeeze into RAM.
Here's the thing: not all the experts need to be in memory at the same time.
With Edge0-35B, we keep expert weights in local storage and bring in the ones needed for inference.
We also use a trained prerouter to predict which experts are needed next, so we can start loading them while the model is still working on the current token.
Still a 35B model. Just a smarter way to handle its memory.
Here's the thing: not all the experts need to be in memory at the same time.
With Edge0-35B, we keep expert weights in local storage and bring in the ones needed for inference.
We also use a trained prerouter to predict which experts are needed next, so we can start loading them while the model is still working on the current token.
Still a 35B model. Just a smarter way to handle its memory.
1
2
24
1,307
A robot doesn't need to write a 2,000-word essay while avoiding a chair.
Smart glasses don't need to run a massive model just to understand “turn up the volume.”
And your phone doesn't need a cloud-sized setup for every AI task.
We talk a lot about how big a model can run on a d talk more about what that device actually needs the model to do.
Smart glasses don't need to run a massive model just to understand “turn up the volume.”
And your phone doesn't need a cloud-sized setup for every AI task.
We talk a lot about how big a model can run on a d talk more about what that device actually needs the model to do.
2
0
8
425
Airplane mode has always meant giving up a few things.
Local AI doesn't have to be one of them.
Once Edge0-35B is downloaded and set up, you can disconnect from the internet and keep chatting with the model.
A small change in your settings. A pretty big change in how AI works.
Local AI doesn't have to be one of them.
Once Edge0-35B is downloaded and set up, you can disconnect from the internet and keep chatting with the model.
A small change in your settings. A pretty big change in how AI works.
1
0
5
276
Our open-source inference engine supports iOS, Android, macOS, and Windows.
github.com/Edge0-AI/Edge0
github.com/Edge0-AI/Edge0
0
0
3
160
AI that works in airplane mode?
Welcome to on-device AI: the model runs on your phone or computer.
Inference can work offline, avoid cloud round trips, and keep your data on the device.
Here’s where those benefits matter.
Welcome to on-device AI: the model runs on your phone or computer.
Inference can work offline, avoid cloud round trips, and keep your data on the device.
Here’s where those benefits matter.
1
0
0
205
And on-device AI doesn't have to mean small models or expensive hardware.
Our internal test of Edge0-35B on an iPhone 17 Pro Max:
- 70+ tok/s prefill throughput
- 14–16 tok/s decoding throughput
- 2.80 GiB reported memory peak
- 4,096-token default context
Large-model inference on a phone.
Our internal test of Edge0-35B on an iPhone 17 Pro Max:
- 70+ tok/s prefill throughput
- 14–16 tok/s decoding throughput
- 2.80 GiB reported memory peak
- 4,096-token default context
Large-model inference on a phone.
1
0
0
42
Edge0 uses 4-bit quantization and loads expert weights as needed. Our 35B model also runs on a 16 GB M2 Mac.
Intelligence at zero distance for everyone.
Code: github.com/Edge0-AI/Edge0
Discord: discord.com/invite/Tg3W2FE…
Intelligence at zero distance for everyone.
Code: github.com/Edge0-AI/Edge0
Discord: discord.com/invite/Tg3W2FE…
0
0
0
28
Your data doesn't have to leave your device for AI to work.
With local AI, your prompts, files, and model outputs can stay on your own device.
No need to send every request to someone else's servers.
And you don't need expensive hardware to do it.
With local AI, your prompts, files, and model outputs can stay on your own device.
No need to send every request to someone else's servers.
And you don't need expensive hardware to do it.
3
0
4
151
Local AI makes sense for:
- Private documents and notes
- Proprietary source code
- Sensitive internal workflows
- Offline applications
Of course, local doesn't mean risk-free. Device security, application permissions and backups still matter.
- Private documents and notes
- Proprietary source code
- Sensitive internal workflows
- Offline applications
Of course, local doesn't mean risk-free. Device security, application permissions and backups still matter.
1
0
1
20
Edge0 brings open-source 8B and 35B AI models to consumer devices, with runtimes for macOS, iOS, Android and Windows.
Want to keep your AI workflow on-device?
Code and setup: github.com/Edge0-AI/Edge0
Discord: discord.com/invite/Tg3W2FE…
Want to keep your AI workflow on-device?
Code and setup: github.com/Edge0-AI/Edge0
Discord: discord.com/invite/Tg3W2FE…
0
0
1
21