Mistral Timestral: open time-series foundation models that forecast on CPU, no GPU needed
Four open probabilistic forecasting models (5M-209M parameters) that predict unseen time series zero-shot, with the 209M topping GIFT-Eval, TIME and fev-bench.
Why it matters
The smallest model beats rivals 90x its size, and the whole family runs on CPU in milliseconds - forecasting without GPU budgets.
What you could build with it
A small energy or retail analytics team could wrap the 28M model into a demand-forecasting service that runs on existing servers, giving shop owners next-week sales predictions without GPU hosting costs.
Does it hold up?
Promising but unverified beyond vendor numbers: weights and inference code are public and CPU-friendly, yet independent leaderboard confirmation was still pending in launch coverage.
Built with Mistral Timestral
- Mistral launches Timestral, an AI that forecasts numbers without a GPUblog · Independent walkthrough of the four sizes, CPU inference times and the Modified MIT revenue clause.
- Mistral's Timestral: 209 Million Parameters, No GPU, and a Suspiciously Perfect Benchmark Scoremedium · Skeptical review that credits the CPU story while warning the benchmark lead is vendor-reported.
- AI Field Notes #126: Mistral Timestral analysisblog · Digest analysis framing the 28M model against Chronos-2 and noting the technical report had not shipped.
Learn more
First spotted on article: source.
More AI models
Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73NVIDIA Isaac Lab 3.0 GA + Isaac OS open runtime + Agile ONE SN humanoidMajor Isaac platform expansion at GTC Berlin Oct 21: Isaac Lab 3.0 is generally available as the open-source robot…models · JEV 0.73Kimi K3.1 Pro: Moonshot's 2.8T open-weight flagship officially launchesMoonshot AI officially launched Kimi K3.1 Pro on Oct 23, a mixture-of-experts model with 2.8 trillion total / 104…models · JEV 0.72NVIDIA Nemotron 3.5 Retriever collection: Embed 8B/1B, Rerank 8B, Parse 631MNVIDIA's Oct 21 collection drop for agentic retrieval: Nemotron-3.5-Embed in 8B BF16, 1B BF16 and 1B NVFP4…models · JEV 0.72NVIDIA Cosmos 3.5 world models: now openly availableNVIDIA's Cosmos 3.5 family of open world foundation models for robots, autonomous vehicles and vision agents is now…models · JEV 0.71NVIDIA RFD4-Proteina: open protein-design BioNeMo modelOpen protein-design model from NVIDIA and the Baker Lab at UW Medicine's Institute for Protein Design, announced at…models · JEV 0.7
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.