{"paper_id":"2508.02739","title":"Kronos: A Foundation Model for the Language of Financial Markets","authors":["Yu Shi","Zongliang Fu","Shuo Chen","Bohan Zhao","Wei Xu","Changshui Zhang","Jian Li"],"paper_date":"2025-08-02","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2508.02739v1","page_url":"https://thequant.space/flowcharts/kronos-a-foundation-model-for-the-language-of-financial-mar/","scores":{"empirical_rigor":9,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes Kronos, a unified, scalable pre‑training framework for financial K‑line data that includes a specialized tokenizer and autoregressive Transformer.","quote":"we propose Kronos, a unified, scalable pre-training framework tailored to financial K-line modeling. Kronos introduces a specialized tokenizer that discretizes continuous market information into token sequences, preserving both price dynamics and trade activity patterns.","summary":"The success of large-scale pre-training paradigm, exemplified by Large Language Models (LLMs), has inspired the development of Time Series Foundation Models (TSFMs). However, their application to financial candlestick (K-line) data remains limited, often underperforming non-pre-trained…","strategy":"Kronos"},"assumptions":{"status":"available","items":[{"assumption":"The data cleaning pipeline can reliably filter out low‑quality K‑line segments such as abnormal price spikes or prolonged inactivity.","type":"data","stated":"explicit","quote":"To guarantee data quality, we develop a streamlined data-cleaning pipeline tailored to the unique characteristics of financial K-line data, which identifies and filters out low-quality segments such as those with abnormal price spikes or prolonged periods of inactivity."},{"assumption":"Trading simulations assume a top‑k/drop‑n portfolio construction with a minimum holding period of 5 days and a limit on daily turnover.","type":"execution","stated":"explicit","quote":"... a minimum holding period of 5 days is enforced for all positions."},{"assumption":"The model can learn hierarchical market dynamics via a coarse‑to‑fine token prediction scheme.","type":"model","stated":"explicit","quote":"This coarse-to-fine prediction scheme allows Kronos to explicitly model multi-scale market dynamics."},{"assumption":"Cross‑attention uses the embedding of ˆbc_t as query and the history h_t as key and value.","type":"model","stated":"explicit","quote":"We use a cross-attention mechanism where the embedding of ˆbc_t acts as the query, and the history h_t provides the key and value."},{"assumption":"The pre‑training corpus consists exclusively of financial K‑line data, assuming this domain‑specific data is sufficient for learning useful representations.","type":"data","stated":"explicit","quote":"To address this fundamental gap in pre‑training, we introduce Kronos, a foundation model built from the ground up on a massive corpus composed exclusively of financial K-line data."}]},"datasets":{"status":"available","items":[{"name":"K‑line dataset","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"45 global exchanges, 7 sampling frequencies","quote":"Our dataset spans over 12 billion observations across 7 sampling frequencies, encompassing a broad spectrum of asset classes drawn from 45 global exchanges."},{"name":"Stock data","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"nine global stock exchanges (in‑distribution and out‑of‑distribution)","quote":"To test both in‑distribution and out‑of‑distribution generalization, we use data from nine global stock exchanges."},{"name":"Binance cryptocurrency data","provider":"Binance","kind":"market-prices","access":"public","period":null,"universe":"All spot trading pairs","quote":"Cryptocurrency: All spot trading pairs available on the Binance exchange.","known_id":"binance"},{"name":"Forex data","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"over 1,000 foreign exchange pairs","quote":"Forex: A comprehensive dataset of over 1,000 foreign exchange pairs."}],"known":[{"id":"qlib","name":"Qlib datasets","provider":"Microsoft","kind":"market-prices","access":"public","quote":"Data Our empirical analysis utilizes daily market data for the Chinese A-share market, sourced from the Qlib plat-form (Yang et al. 2020), an open-source framework for quantitative finance.","mentions":3,"used_in":2}],"sample":{"start":null,"end":null,"market":"Chinese A-shares","asset_class":"equities","frequency":"daily"},"data_resolution":"intraday-bars"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":false,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"no","evidence_card":"partial","evidence_card_quoted":true},"conflict_kind":"graded","card_detail":"The paper discusses turnover limits and a maximum number of stocks bought/sold daily but does not specify explicit cost numbers.","card_quote":"To manage turnover and trading costs, a maximum of n stocks are bought or sold daily, and a minimum holding period of 5 days is enforced for all positions.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"Kronos improves price series forecasting RankIC by 93% over the leading TSFM and 87% over the best non‑pre‑trained baseline.","kind":"performance","quote":"Kronos excels in a zero-shot setting across a diverse set of financial tasks. On benchmark datasets, Kronos boosts price series forecasting RankIC by 93% over the leading TSFM and 87% over the best non-pre-trained baseline."},{"finding":"Kronos achieves a 9% lower MAE in volatility forecasting compared to baselines.","kind":"predictive-accuracy","quote":"It also achieves a 9% lower MAE in volatility forecasting."},{"finding":"Kronos improves generative fidelity for synthetic K‑line sequences by 22%.","kind":"performance","quote":"and a 22% improvement in generative fidelity for synthetic K-line sequences."},{"finding":"In backtesting simulations, Kronos demonstrates superior investment performance across multiple asset classes.","kind":"performance","quote":"Subfigure (e) presents the investment simulation backtesting results."}],"reported_performance":{"strategy":"Kronos","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":67,"citations":{"cited_by_in_corpus":6,"by_independent_authors":5,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":2,"unclassified":3},"independent_papers":[{"paper":{"paper_id":"2611.19937","title":"AllocFlow: End-to-End Portfolio Trading Action Generation from Raw Market Data","paper_date":"2026-11-15","authors":["Annan Wang","Yuchen Xie","Yu Cen","Leon Liangyu Chen","Qianyu Julie Zhu","Xin Li"],"empirical_rigor":8,"page_url":"https://thequant.space/flowcharts/allocflow-end-to-end-portfolio-trading-action-generation-from-raw-market-data/"},"how":"uses_method_or_data","quote":null}],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":true,"code_url":"https://github.com/shiyu-coder/Kronos","note":null},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T17:30:53Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}