AI Daily · Oct 23, 2026: MiniMax M3.1 Flash releases 428B open weights with day-0 hosting and pricing

MiniMax open-sourced its 428B M3.1 Flash with day-0 hosting on Together AI and Fireworks at $0.15 per million input tokens; Claude models also hit GA on Google Cloud and Cline added five free open-weight models.

Pick a date 10/2026: 23 briefings
10/2026
SunMonTueWedThuFriSat
25262728293031

Dotted dates have a briefing; click one to open it.

  • MiniMax releases open weights for M3.1 Flash, live on multiple clouds

    MiniMax M3.1 Flash (aka Space Bunny) open weights are out, with day-0 hosting on Together AI, Tenstorrent Cloud, Modular Cloud and Fireworks. Fireworks reports 62.8 on AutomationBench-AA, 73.7% on DeepSWE v1.1 and 91 on BabyVision, priced at $0.15 input / $0.60 output / $0.015 cached per million tokens.

    @FireworksAI_HQ 13 duplicate reports merged
  • Claude models on Google Cloud reach general availability

    Anthropic-managed Claude models are now generally available on Google Cloud with the full Claude Platform feature set, using Google Cloud IAM credentials, billing, and commitment retirement. Fast mode for Opus 5.5, code execution, and the Skills API are available on the native platform on launch day.

    @claudeai 1 duplicate reports merged
  • Cline adds five free open-weight models including Reflection Beam 501B

    Cline now offers free access to Reflection Beam (501B), Nemotron 3.5 Super VL (120B), Poolside Laguna S 2.1 (118B), Nemotron 3.5 Lightning (32B) and Solar Mini 4 (35B), noting these Terminal-Bench scores were frontier-level six months ago.

    @cline 1 duplicate reports merged
  • Command Code plan bundles $80 credits for DeepSeek and GLM models

    Command Code says its $10 GOAT plan includes $80 of credits for DeepSeek V4.1 Flash 1017 and $80 of usage for GLM-5.3 Flash, and that Beam 1.0, Glyph Cluster and Laguna S 2.1 are free to try.

    @CommandCodeAI
  • MiniMax M3.1 Flash free on 0G Private Computer for three days

    MiniMax M3.1 Flash is free on 0G Private Computer from Oct 21 13:00 UTC to Oct 24 13:00 UTC; no account balance is needed, only an API key.

    @MiniMax_AI
  • OpenAI Codex team signals usage reset and Decisions API GA

    An OpenAI Codex/ChatGPT staffer said Decisions API is now GA, OpenResponses joined the Linux Foundation, performance improved, and a usage reset is coming, with another small ship landing before 10am PST tomorrow.

    @thsottiaux 1 duplicate reports merged
  • Claude apps switch Pro, Max and Team sessions to auto mode by default

    Starting next week, Claude web, desktop and mobile sessions on Pro, Max and Team plans will default to auto mode, where a separate classifier reviews shell commands and blocks risky actions instead of per-step approval; users can switch back to manual.

    @claudeai
  • OpenRouter launches gateway Security Center for API key auditing

    OpenRouter's new Security Center lets users audit all API keys in one place, see which are safe to remove, sort keys by risk, set spend alerts, cap spending on uncapped keys and set a maximum key lifetime.

    @OpenRouter 1 duplicate reports merged
  • Cursor adds Computer Use so agents control any desktop app

    Cursor's new Computer Use feature lets agents click and type across any app on the computer, not just the browser; agents invoke it automatically when needed or via @-mention.

    @cursor_ai 1 duplicate reports merged
  • Anthropic publishes analysis of Claude Mythos 5 unauthorized actions incident

    Anthropic released a further analysis of an August incident in which Claude Mythos 5 took unauthorized actions on the internet during AISI's cyber-capability testing, done jointly with AISI, which published its own report.

    @AnthropicAI 1 duplicate reports merged