Someone finally built GPT for candlesticks. Kronos, from developer shiyu-coder, treats K-line charts (OHLCV bars) as a language: a dedicated tokenizer quantizes continuous price and volume data into discrete tokens, then a decoder-only Transformer predicts what comes next. Same recipe as an LLM, pointed at markets instead of text.
Pre-trained on 45+ global exchanges, MIT-licensed, and already accepted to AAAI 2026. The model family runs from a tiny 4.1M-parameter version up to 499.2M, so you can pick your compute budget.
What you can actually do with it
Weights ship on Hugging Face (NeoQuasar/Kronos-base and friends), so you load them locally for inference — no API key, no rate limits. There’s a live demo forecasting BTC/USDT, a full fine-tuning pipeline for your own instruments, and Qlib integration to plug straight into a quant backtest.
Why it’s trending
It’s the first open foundation model built specifically for market data, and both the quant crowd and the AI crowd noticed at once. Up ~400 GitHub stars in a single day, one of the fastest-climbing repos this week.
You Might Also Like
- Glm 5 Just Dropped and its the Open Source Model Nobody saw Coming
- Evo 2 Just Dropped and it Might be the Most Important Open Source Model Youve Never Heard of
- Nvidia Nemotron 3 Super 120b Parameters 12b Active the Math Behind the Fastest Open Source Reasoning Model
- Tsinghua Tested an ai Classroom on 500 Students now Openmaic is Open Source and Trending
- Onyx Hits 19 7k Github Stars the Open Source Answer to Gleans 7 2b Enterprise ai Play
Top AI Product 
Leave a comment