UpdatedWeights are listed: 23B active does not make this a 23B deployment
Follow the public shards, 428B total/23B active, serving recipe and community license, separating model use from self-hosting.
Creations
NewEditor’s pick · Models and editable work
Kimi K3.1 Pro: what changes beyond 55?
Connect the official 3D entry and CAD evaluation to model access, Code tiers and speed quotas. Distinguish the image’s K3 High label from a new-model result.
Explore the work and processWorth exploring now
Explore a change, a piece of work or an experience, and the possibilities it opens up.
Updated
Topic · UpdatedDecisions is generally available: routing, probability and severity are different questions
Follow Decisions GA through probability, categories, severity, refusals and actions, with Jev versions, thresholds and costs kept within a matched task.
Topic · Updated1Password: why approving a sign-in still leaves a purchase decision
Use 1Password’s Muse illustration to separate sign-in, account actions and purchases, then check each partner’s rollout.
NewFlint: what does a coding agent need to work on your laptop?
Flint offers another local route. Connect full coding tests, long-context speed and hardware to understand how model, engine and tools work together.
Follow a question further
Connect capabilities and tools with human choices and real experiences. Return to the latest addition when new work or understanding emerges.
Topic · UpdatedDeepSeek V4.1 Flash: from images to projects with web tools
Connect images, files and web tools with 1017 evaluations and the DeepInfer engine. Distinguish model capability, service deployment and real tasks; the latest addition explains offline scope and remaining online components.
Latest additionWith DeepInfer open, what remains between an engine and a service?
Topic · UpdatedOpenAI DevDay 2026: how AI takes on ongoing work
How do dots, Space, Pages and cloud Codex connect ongoing work? Start with the 3D venue demo and product interfaces, then examine model costs, developer tools and the subscription you already use.
Latest additionSlides: shared editing and export review
Topic · UpdatedMeta models: from FAIR and Llama to Wonder
Wonder Full enters the current frontier capability range. Follow Meta’s research, open models, team transitions and original leadership interviews to connect this change to works, tools and everyday life.
Latest additionPrometheus: how Wonder training keeps moving
Topic · FollowingMeta Muse: from personal agent to portable hardware and everyday tasks
Follow Spark and Wonder, shared-task tests and developer integration through Charm, glasses and everyday tasks to see how an agent remembers context and continues work.
Latest additionHow do products gathered across stores become a completed purchase?
Topic · FollowingOpus 5.5: making short films and a full music video
Connect capabilities, independent tests and real creation through two short films and a full hand-painted music video. Follow the Opus–Sonnet handoff, continuity checks and production costs.
Latest additionThe same Eiffel Tower brief: how do we check the interaction?
Topic · FollowingWhat changes with Astra: task comparisons and real work
Examine Astra through shared drawings, handwritten edits, furniture errors and a tax workbook, then explore editable scenes, family memories, interactive websites and video editing.
Latest additionA newsletter index while cooking: what changed before launch?
Topic · FollowingGPT Image 2.5: photos, styles and edits
Explore photo remixes, poster styles, focused edits and sketches—and how they fit everyday creative work.
Latest additionHow do separate person, dress and bag references become one image?
Topic · FollowingMicroDuck: how a little robot learns new moves
Connect the physical robot, simulation and open software: what has been demonstrated, what the preorder includes and which follow-ups matter.
Latest additionHow can you arrange four parts for a robot-duck choir?