Start free trial

TRUE, BUTSelf-Marked

Weights are out, with conditions

But one benchmark sits close to Opus 5.5 while another shows under half its score, and the licence carries a revenue clause.

“Kimi K3.1 Pro's model weights are out, as an open-weight model that nears top proprietary models.”

Hacker News post, 4 November 2026, and Moonshot AI's model card

Recorded 5 November 2026. Our ruling means: literally true, materially misleading.

Read the story →

Receipts pack

The evidence file for this verdict

25 receipts: 6 support the claim, 1 contradict it, 18 add context. 3 figures on the record, 2 traced to a quoted receipt.

The receipts

Each source, the line we quote from it, its date and which way it cuts. Free to read and to cite.

  1. 1Supports the claimhuggingface.co ↗

    Kimi K3.1 Pro is an open-weight, natively multimodal agentic model

    Look up an archived copy ↗

  2. 2Supports the claimhuggingface.co ↗

    Terminal-Bench 4.0 65.7 57.9 58.2 66.4 42.3 41.8

    Look up an archived copy ↗

  3. 3Contradicts the claimhuggingface.co ↗

    Program Bench 31.5 53.5 50.0 65.0 29.1 19.0

    Look up an archived copy ↗

  4. 4Contexthuggingface.co ↗

    the aggregate revenue of the Licensee and its affiliates exceeds 20 million

    Look up an archived copy ↗

  5. 5Contexthuggingface.co ↗

    must enter into a separate agreement with Moonshot AI

    Look up an archived copy ↗

  6. 6Contexthuggingface.co ↗

    "Kimi K3.1-Pro" must be prominently displayed on the user interface

    Look up an archived copy ↗

  7. 7Contexthuggingface.co ↗

    The requirements set forth in Sections 2 and 3 do not apply to: (a) internal use of the Software

    Look up an archived copy ↗

  8. 8Contexthuggingface.co ↗

    Permission is hereby granted, free of charge

    Look up an archived copy ↗

  9. 9Contexthuggingface.co ↗

    in total over any consecutive 12 months

    Look up an archived copy ↗

  10. 10Contexthuggingface.co ↗

    commercial products or services that have more than 100 million monthly active users

    Look up an archived copy ↗

  11. 11Contexthuggingface.co ↗

    operates a Model as a Service business

    Look up an archived copy ↗

  12. 12Supports the claimvktr.com ↗

    Kimi K3.1 Pro Nears Top Proprietary Models as Moonshot Limits Its Cyber Skills

    Look up an archived copy ↗

  13. 13Contextvktr.com ↗

    It traced the trend to cyber content embedded in the general mid-training corpus and to cyber tasks in its RL environments, then removed both for the open-weight release

    Look up an archived copy ↗

  14. 14Contextvktr.com ↗

    On xdev-bench-hard, Moonshot's internal suite of exploit tasks

    Look up an archived copy ↗

  15. 15Contextvktr.com ↗

    Its self-reported results show a model that leads in some areas and trails noticeably in others

    Look up an archived copy ↗

  16. 16Supports the claimhuggingface.co ↗

    SWE Marathon v1.1 51.7 45.6 42.5 46.3 40.0 42.5

    Look up an archived copy ↗

  17. 17Contextvals.ai ↗

    It ranks #9 of 33 models on the Vals Index with 62.52%.

    Look up an archived copy ↗

  18. 18Supports the claimvals.ai ↗

    Its best result is #3 of 60 on Terminal-Bench 4.0

    Look up an archived copy ↗

  19. 19Contextvktr.com ↗

    The released model solved 81%, up from 33.3% for K3.

    Look up an archived copy ↗

  20. 20Contextvktr.com ↗

    an unmitigated checkpoint solved 90.5% of problems

    Look up an archived copy ↗

  21. 21Supports the claimvktr.com ↗

    ahead of GPT-6 Astra (58.2%) and narrowly behind Claude Opus 5.5 (66.4%)

    Look up an archived copy ↗

  22. 22Contextvktr.com ↗

    Several of the evaluations Moonshot cites are internal benchmarks that have not been independently verified.

    Look up an archived copy ↗

  23. 23Contextvktr.com ↗

    not to release open weights that would extend the absolute frontier of cyber capability

    Look up an archived copy ↗

  24. 24Contextvktr.com ↗

    It also plans to gradually raise the cyber capability available in open weights

    Look up an archived copy ↗

  25. 25Contexthuggingface.co ↗

    Benchmark Kimi K3.1 Pro (max) Claude Fable 5.1 (max, w/ fallback) GPT-6 Astra (max, w/ fallback) Claude Opus 5.5 (max, w/ fallback) DeepSeek-V4.1-Flash-1017 (max) GLM-5.3 (max)

    Look up an archived copy ↗

Members' extras

What each receipt proves, and what would change our mind.

  • What each receipt above proves, in our words
  • The figures table, each number traced to the receipt that states it
  • The trick behind the claim, explained, with the guide to spotting it
  • What would change our verdict
  • Our call on the record, with its due date

The verdict, the opening and every receipt above stay free. Membership adds our analysis to the evidence file for every checked story.

First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge.