AIToday
Large Language ModelsAI Coding AssistantsOpen-Source AIITmedia AI+Published: Oct 26, 2026, 13:01 JST

Moonshot AI unveils Kimi K3.1 Pro, claiming edge over GPT-6 Astra

Send on LINE
Moonshot AI unveils Kimi K3.1 Pro, claiming edge over GPT-6 Astra

3 Key Points

  1. What happened

    Moonshot AI announced Kimi K3.1 Pro on October 23, an open-weights model with 2.8 trillion total parameters, a 1 million-token context window, and a 65.7% score on Terminal-Bench 4.0, above GPT-6 Astra and close to Claude Opus 5.5.

  2. Why it matters

    The model's benchmark results suggest Moonshot AI can field an openly available system competitive with leading proprietary models on coding tasks, potentially giving developers a free alternative to closed APIs.

  3. What to watch

    The model weights are scheduled to be released by November 4.

WHO IT HITSDevelopers and companies building on open-weights models may gain a new high-performance option, though they will need to weigh its stated exclusion of cybersecurity-related training data when considering security-sensitive applications.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Moonshot AI's Kimi K3.1 Pro is positioned as an upgrade to the prior Kimi K3, with post-training applied to strengthen coding, 3D modeling, and financial tasks. The company is offering a Highspeed variant that runs 2 to 3 times faster and is available from the same day on Kimi.ai, its app, and via API. The API pricing was lowered from Kimi K3, though Artificial Analysis estimates the cost per task is nearly the same at $1.98 due to increased output token counts. The open-weights release, planned by November 4, will be trained without cybersecurity-related data, a measure Moonshot AI says is intended to prevent misuse. If the weights are released as planned, the model could become the top open-weights model on Artificial Analysis's Intelligence Index, where it currently ranks fourth overall with 55 points.

FAQ
How much does the API cost?
The API is priced at $2 per 1 million input tokens and $10 per 1 million output tokens, a reduction from Kimi K3. The Highspeed version costs twice the standard version's API price.
What is different about the open-weights version?
The open-weights version was trained with cybersecurity-related data removed to prevent misuse. The weights will be released by November 4.
What is the context window size?
Kimi K3.1 Pro has a maximum context length of 1 million tokens.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleMeta falls 7% as $34.9 billion capex outruns cash flow