Kronos

# Fix list: Kronos

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/kronos, the October 2026 research run, assessed 25 October 2026. Grade C, 55.2 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Kronos: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reach, 55 out of 100, up to 17.3 more on the total

Why it scored 55: Memory 30, the publisher's Kronos-mini with its tokeniser is 0.03 GB and even Kronos-base is 0.43 GB in FP32. Runtimes 0, it runs only through the publisher's own PyTorch package from GitHub, and none of the six named runtimes supports it. Builds 0, FP32 safetensors only. Terms 15, MIT with nothing added. Download 10, nothing is gated. Hosted 0, no inference provider or publisher API serves it.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reach):

Whether an agent can run it, and where.

- 0 to 30, memory. The smallest build from the publisher or a runtime's own library that keeps 4-bit precision or better, at its file size plus a fifth for the runtime and a working context. 30 for 16 GB or less, 25 for 24 GB, 20 for 32 GB, 15 for 64 GB, 10 for what fits our reference machine (the GPU share of 128 GB of unified memory), 5 for 256 GB, 0 above.
- 0 to 25, runtimes. 5 for each of llama.cpp, Ollama, vLLM, SGLang, LM Studio and Hugging Face Transformers that runs the architecture in a released version, from the runtime's own code, library or documentation. A publisher's fork or an unreleased branch doesn't count. 25 at most.
- 0 to 15, builds from the publisher. 15 for quantised builds at two precisions or more (FP8, 4-bit, GGUF), 10 for one, 0 when only other people's quantisations exist.
- 0 to 15, terms of use. 15 for an OSI-approved licence with nothing added, 10 for a custom licence that lets anyone use it commercially for free below a threshold or under a use policy, 5 for research or non-commercial use only, 0 for no licence or unclear terms.
- 0 to 10, download. 10 when the files download with no account and no terms to accept, 5 for a click-through that needs an account, 0 when a person approves each request.
- 5, hosted too. The publisher's own API serves it, or at least two inference providers do, so an agent without the hardware can still call it.

When our reference-machine runs start, speed on that machine (tokens per second and time to first token at the build that fits) joins Reach.

## 2. Risk, 50 out of 100, up to 15.4 more on the total

Why it scored 50: File safety 15, safetensors only, but the only documented loading path is the publisher's Python package in place of a runtime. Provenance 15, the files are in the NeoQuasar account the publisher's README links, with SHA-256 hashes. Safety evaluation 0, none published. Agent guidance 5, the README's general cautions that the fine-tuning pipeline isn't a production trading system. Licence risk 15, plain MIT; the broken licence link is charged to file health. Security programme 0, no reporting route found.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-risk):

What could go wrong when you run it, and what the publisher has done about it. A higher score means less risk.

- 0 to 25, file safety. 25 for weights in safetensors or GGUF only, 15 when pickle files (.bin, .pt, .pth) sit beside them, 0 for pickle only. 10 off when the documented way to load it or to build its prompts runs code from the repository (`trust_remote_code`, or the publisher's scripts in place of a template).
- 0 to 15, provenance. 10 when the files come from the publisher's own organisation, the one its site or docs link to, and 5 for a checksum on every file.
- 0 to 20, safety evaluation. 20 when results are published and someone other than the publisher ran them or can rerun them, 10 for the publisher's own results, 0 for none.
- 0 to 15, guidance for agent use. 15 for documented prompt-injection guidance or an instruction hierarchy the model was trained to follow, 10 for a statement of what its trained refusals cover and the safeguards to add around it, 5 for general cautions, 0 for none.
- 0 to 15, licence risk. 15 to start, 5 off for each of these, a use policy or field-of-use limits beyond the law, terms the licensor can change or revoke, a revenue or user threshold, and a licence that changed without a dated note or whose link is broken. Not below 0.
- 0 to 10, a security programme. 5 for a way to report a flaw (security.txt, a disclosure policy or a security contact), 5 for advisories handled in public.

When our reference-machine runs start, measured robustness (prompt injection with AgentDojo, over-refusal with XSTest) joins Risk.

## 3. Agent interface, 60 out of 100, up to 6.2 more on the total

Why it scored 60: Read for a forecasting model, the tool-calling, trained tool calling and structured-output lines stand for what it takes and returns. The card documents the DataFrame and timestamps it takes and the DataFrame it returns, and the publisher's KronosPredictor builds and parses them (25). It was trained on six K-line fields with calendar fields, one series per forecast (10). Output is a typed DataFrame indexed by the requested timestamps, the mean of the sampled paths (15). The input format is given only as code (10). Reasoning control and context score 0, with no reasoning and a 512-bar context (2,048 for Kronos-mini).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-interface):

What a model and a developer need to talk to it. Speech, image and other models with no chat interface read the tool-calling and structured-output lines for what the model takes and returns (voices, timestamps, streaming, typed outputs), and the note says how.

- 0 to 20, the chat template. 20 for a template in the repository (Jinja, in the tokenizer config or its own file), 10 when the prompt format is only given as code you run.
- 0 to 25, the tool-calling format. 25 when the publisher documents how tools are declared and how calls come back, with the parser to use in a named runtime, 10 when only the template or a runtime's own recipe says it.
- 0 to 20, tool calling the model was trained for. 20 with several calls in one turn, 10 for one call at a time.
- 0 to 15, structured output. 15 when JSON or JSON Schema output is documented for the model with how to ask for it, 5 when the card claims it without saying how.
- 10, reasoning control. Thinking that can be turned off or budgeted per request, with where the reasoning text comes back.
- 0 to 10, context. 10 for 128K tokens or more trained, 5 for 32K to 127K. Context reached only by scaling the position encoding counts one band down.

## 4. Documentation and upkeep, 61 out of 100, up to 6 more on the total

Why it scored 61: Model card 11, intended use, architecture, parameters and context are there, limitations, languages and a cut-off aren't. Training data 20, the paper lists every exchange with asset and observation counts, the filtering and a June 2024 end. Last change 5, the reference code on 13 April 2026 and the files in mid-2025. Responsiveness 20, the publisher answered 10 of 18 Hugging Face discussions, though none since November 2025. Version policy 5, a first release. File health 0, the cards point to example data the repository has deleted and to licence and requirements links on a branch that doesn't exist.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-upkeep):

- 0 to 20, the model card. Intended use, limitations, architecture, parameters, context length, languages and knowledge cut-off. 20 for all seven, 3 off for each that's missing.
- 0 to 20, training data. 20 for the data published or listed with sources and sizes, 15 for a data card that names the sources or the mix, the cut-off and the filtering, 10 for a general description with a cut-off, 5 without one, 0 for none.
- 0 to 20, the last change to the files, the template or the reference code. 20 within 30 days, 15 within 90, 10 within 180, 5 within a year, 0 older.
- 0 to 20, responsiveness. The publisher's replies to the 40 most recent questions and reports in the repository's discussions or issues. 20 for most, 10 for some, 5 for a few, 0 for none.
- 0 to 10, version policy. 10 for earlier versions kept in their own repositories with a dated note when one is replaced, 5 for earlier versions kept without notes or a first release, 0 when files were replaced in place without a note.
- 0 to 10, file health. 10 when config, tokenizer, template and card agree and no report of a broken file waits without a reply, 5 when one does, 0 when the files contradict the card.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Capability is pending. The checklist has no forecasting or time-series area, so a forecasting benchmark couldn't count however it was run, and four searches on 25 October 2026 found no independent benchmark of Kronos with a published method (two 2026 papers benchmarking time-series models on financial data don't include it); the total is capped at 61
- Unchecked, GitHub issues on shiyu-coder/Kronos (213 open), since the GitHub API isn't enabled for this session and the issues page was refused; responsiveness rests on the 18 Hugging Face discussions
- Unchecked, github.com/robots.txt, which our shell proxy and WebFetch both refused; the repository page was read once by WebFetch and the code by git clone
- The cards' licence link is read as broken from raw.githubusercontent.com (404 on main, 200 on master); github.com's own blob page answered 403, so a redirect there wasn't ruled out
- File health reads the deleted sample data as the files contradicting the card; a gentler reading of it as one disagreement gives 5 and Upkeep 66
- Whether the 13 April 2026 merges by the publisher of contributors' pull requests count as the publisher's own change; they're read as such, since the publisher merged them into master

## Weaknesses

- Runs only through the publisher's own PyTorch code from GitHub, in no named runtime and with no quantised build
- A 512-bar context for the small and base sizes, and the mean forecast only, without quantiles
- The sample data the quickstart reads was deleted on 13 April 2026, and the cards' licence link points at a missing branch
- No safety results, no security contact and no publisher reply in its Hugging Face discussions since November 2025
- The publisher's forecasting results come from its own paper, with no independent benchmark found

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Clone https://github.com/shiyu-coder/Kronos and `pip install -r requirements.txt`, then load `Kronos.from_pretrained("NeoQuasar/Kronos-small")` and `KronosTokenizer.from_pretrained("NeoQuasar/Kronos-Tokenizer-base")` and wrap them in KronosPredictor
- Pair Kronos-mini with Kronos-Tokenizer-2k and Kronos-small or Kronos-base with Kronos-Tokenizer-base, the cards' pairing
- The examples read examples/data/XSHG_5min_600977.csv, which no longer exists; point them at tests/data/regression_input.csv, 2,500 five-minute bars
- Keep the lookback at or under max_context, 512 bars, and pass the future timestamps you want as y_timestamp, since the calendar fields come from them
- Raise sample_count for a steadier forecast, but expect only the mean back; call the lower-level auto_regressive_inference yourself if you need the paths
- Treat forecasts as model output to check, not signals; the README says its own backtest pipeline isn't a production trading system

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.