Ai2 releases HiRO-ACEv2, generating global 3 km extreme-weather simulations at about 1/300th of the cost
A 100 km climate emulator plus a downscaling model that refines its output to 3 km, with weights on Hugging Face.
// Key points
- ACE runs at 100 km and the new downscaling model HiROv2 refines its output to 3 km, predicting several high-resolution fields at once rather than only rainfall.
- Generating global 3 km data costs about 1/300th of running X-SHiELD, a high-resolution model that explicitly simulates storms; for a single region the savings can exceed 100x.
- Given coarsened X-SHiELD data, HiROv2 matches the target’s high-resolution statistics within 3%, including events that occur about once in a million data points; a streamlined version that could be more than 10x faster is in progress.
Builder's takeI don’t work on weather, but the shape is worth borrowing: a cheap coarse model plus a dedicated refinement model, paying for detail only where it’s needed. Read alongside this week’s small-model releases, cheap coverage first and a specialist model for precision second is becoming a general way to cut cost.
// Background · from #Open models
Full timeline →- Oct 15 Microsoft releases UserRepGen-4B under MIT: it turns multi-turn conversations into user descriptions for simulators, scoring 4.95 on fidelity against GPT-5.5’s 3.90
- Oct 14 Liquid AI releases open-weight LFM2.5-24B-A2B, a reasoning agent model with 2B active parameters that decodes 184 tokens/s on an M5 Max
- Oct 14 NVIDIA lines up software for RTX Spark laptops shipping October 16, including the Nemotron 3.5 Bolt 74B agent model on October 15