Introducing Edge0-145B: Frontier AI Capabilities on Your Own Computer

Edge0 open-sourced Edge0-145B, billed as the first open-source on-device model with 145 billion parameters, alongside the smaller Edge0-120B, a unified inference framework and deployment tools. The company said the model runs locally on a compatible laptop such as a MacBook Pro with 29GB peak memory, prefill above 1,100 tokens/s, decode above 80 tokens/s, roughly 120W typical power draw and 128K–256K context, and that its coding, reasoning and agent results are competitive with specific frontier cloud models. Edge0 pitched the release as removing the tradeoff between on-device capability and hardware limits, letting developers plan, write and debug code and check results visually without cloud token or API fees, data-handling constraints or connectivity dependence, though hardware and electricity costs remain. The release follows Edge0's earlier Edge0-35B phone model, which reported sub-2.5GB peak memory and 37+ tokens/s on an iPhone 18 Pro Max. Next, watch independent benchmarks of the 145B model's frontier-parity claims and how quickly developers adopt the open weights and framework.