Skip to content
NewA calmer, faster ZeroHour
StoryModel releaseinfo

Moonshot ships Kimi K3.1 Pro; full weights still due November 4

7 sources · 8 articlesFirst reported Updated ()
What's new

October 25: Interconnects hosts say weights follow about two weeks of exclusive access, cite mixed quality and a roughly 20-point Terminal-Bench Science gap versus GPT-6 Astra, and call a paid-promotion screenshot uncorroborated.

Merged summary · x-ai/grok-4.7 · rewritten as coverage arrivesSummaries by Vani AI

Kimi K3.1 Pro is live on Moonshot and reseller APIs, while full weights remain promised for November 4.

Moonshot AI released Kimi K3.1 Pro, a post-trained mixture-of-experts model it describes as 2.8 trillion parameters with 104 billion active and a 1-million-token context, plus a Highspeed tier said to be 2–3× faster at twice the price except cached input. Standard Moonshot pricing is $2 per million input tokens and $10 output. The Decoder says Artificial Analysis scores it 55, three points behind Claude Opus 5.5 at 58. Simon Willison, The Decoder, and Latent Space say weights stay unpublished until November 4, 2026; Hacker News and MarkTechPost describe the model as open-weight now. Vani lists a cheaper API and, on October 25, a Soft-Derisked variant; Moonshot and Vani cite different hallucination drops.

Key facts
  • Highspeed is 2–3× faster and costs twice as much, except cached input.
  • Moonshot reports 65.7% on Terminal-Bench 4.0; The Decoder says K3 scored 13%.
  • After cyber-data removal, Moonshot says xdev-bench-hard fell from 90.5% to 81.0%.
  • Vani API is $1.40 input and $7 output; Soft-Derisked is $5 input and $7 output.
  • Hallucination claims: 23.4% to 8.3% (Moonshot) versus 23.35% to 6.85% (Vani).
  • Latent Space: eighth in Code Arena WebDev at 1,692; Cyber Index 64%, second.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

Timeline

dated events in the reporting · oldest first
  1. Oct 23, 2026Kimi K3.1 Pro and Highspeed go live on Kimi products. · Hacker News
  2. Oct 24, 2026Vani lists Kimi K3.1 Pro on chat and its API. · Vani AI
  3. Oct 25, 2026Vani adds a Soft-Derisked K3.1 Pro listing. · Vani AI
  4. Nov 4, 2026Moonshot plans to release the full model weights. · The Decoder

What each outlet adds

beyond the shared facts
  • The DecoderParameter split, price cut, cyber-data cleaning, and AA 55 versus 58.
  • Simon WillisonWeights still closed; Highspeed twice price; parity with GPT-6.1 Sol.
  • Latent SpaceArena and Cyber Index ranks, plus Kimi Hosted Agents.
  • Vani AIReseller pricing and an 18-plus Soft-Derisked API variant.
  • InterconnectsMixed quality impressions and an uncorroborated paid-promotion claim.

Coverage

oldest first · one row per article · UTC
  1. · 2d ago Hacker News · AIFirst report

    Kimi introduces K3.1 Pro, a 3T model with 1M context

    The model and a faster Highspeed variant are live in Kimi products now; full weights are promised by November 4.

  2. · 2d ago The Decoder

    Moonshot releases Kimi K3.1 Pro, fourth on AA index

    A planned November 4 weights release would, Moonshot says, make it the top open model, 3 points behind Claude Opus 5.5.

  3. · 2d ago Simon Willison

    Moonshot puts Kimi K3.1 Pro on the Kimi API

    Weights stay closed until a promised November 4 release; a faster tier costs twice as much.

  4. · 1d ago MarkTechPost

    Moonshot AI releases open-weight Kimi K3.1 Pro

    The title lists 2.8T total parameters, 104B active parameters, and a 1 million-token context.

  5. · 1d ago Latent Space

    Moonshot launches Kimi K3.1 Pro; weights promised November 4

    The 2.8T mixture-of-experts model, with 104B active parameters, scored 55 on Artificial Analysis; the API is live but weights are not.

  6. · 1d ago Vani AI

    Kimi K3.1 Pro is available on Vani Chat and API

    Moonshot says unsupported claims fell from 23.35% to 6.85%, with open weights planned for early November.

  7. · 10h ago Vani AI

    Vani adds Kimi K3.1 Pro Soft-Derisked to chat and API

    API pricing is $5 per million input tokens and $7 for output, with an 18-plus confirmation in chat.

  8. · 8h ago Interconnects

    Podcast weighs mixed reaction to Kimi K3.1 Pro benchmarks

    Hosts say weights are due by November 4 after about two weeks of exclusive access, with quality impressions still mixed.