moonshotai/Kimi-K3

Timeline

1. Oct 19, 2026 · Open-source release · 1 source
   sgl-project/sglang v0.5.22
   | Model | Type | Cookbook |
2. Oct 18, 2026 · Product / feature launch · 1 source
   72 hours, four big open weights drops. The only one that fits in 24GB came from NVIDIA, of all companies.
   Quick recap, all numbers from the official repos and model cards:
3. Oct 17, 2026 · Tutorial / explainer · 14 sources
   Local and Distributed Large LLM Inference with MLX and RDMA on Mac Studio
   In this post, we show how to run the latest frontier-class model — Muse Spark Open, with nearly a trillion parameters — right from your desk, using the latest version of MLX, released today.
4. Oct 17, 2026 · Open-source release · 2 sources
   StepFun Releases Step 5 Preview Open Weights: 600B MoE, 27B Active, 1M Context
   StepFun has released Step 5 Preview as open weights.
5. Oct 17, 2026 · Open-source release · 1 source
   Perplexity Open-Sources SPACE, a Firecracker MicroVM Sandbox for AI Agents
   Perplexity has open-sourced SPACE, the sandbox platform behind every task in Perplexity Computer.
6. Oct 16, 2026 · Opinion / analysis · 1 source
   Kimi K3, three months after the hype: too big for almost anyone to run locally, and smaller models have already caught it
   K3 came out in July hyped as the best open model you could download, and it scored 57 on Artificial Analysis, right up there with Opus 4.8 and GPT-5.5.
7. Oct 16, 2026 · Opinion / analysis · 1 source
   Understanding How Neural Networks Learn (with interactive visualizations)
   This essay rewrites fundamental assumptions about how neural networks learn and connect, with some of the best interactive visualizations I've ever seen.
8. Oct 15, 2026 · Research paper · 1 source
   Can LLMs predict the outcomes of machine learning experiments?
   Recent advances in AI make Recursive Self-Improvement (RSI) increasingly plausible.
9. Oct 15, 2026 · Research paper · 2 sources
   MLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning Engineering
   We propose a post-training method for autonomous long-horizon MLE optimization and apply it to train MLExpert, a 35B Mixture-of-Experts agentic model.
10. Oct 14, 2026 · Research paper · 1 source
    How to Train Your Little Dragon: Teaching a Small Language Model to Construct and Refine GPU Optimization Scripts
    We introduce HOW TO TRAIN YOUR LITTLE DRAGON, a training method that teaches a small language model both to construct complete optimization scripts and to improve existing ones.
11. Oct 13, 2026 · Research paper · 1 source
    Compiling Expert Procedures into Verified Programs: When a Small Model Is Enough
    In procedural tool-use domains, how much of the gap between base models is procedure rather than capability?
12. Oct 13, 2026 · Research paper · 1 source
    Output Discipline Sets the Floor: A Falsification-Guided, Model-Agnostic Agent Protocol for Auditable Hypothesis-Driven Decision Making
    What binds large language models on auditable, hypothesis-driven decision tasks, where every candidate can be checked against the evidence?
13. Oct 13, 2026 · Research paper · 13 sources
    Sharing AI progress in mathematics
    OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.
14. Oct 13, 2026 · Model release · 2 sources
    Z.ai GLM-Legal Beats Claude Fable 5 and Kimi K3 on Harvey LAB
    GLM-Legal is a legal AI model from Z.ai, built on GLM-5.3-Flash and trained for agentic legal work.
15. Oct 12, 2026 · Research paper · 1 source
    ReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding
    NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are...Hybrid models interleave state-space model (SSM) layers with a few attention layers.
16. Oct 12, 2026 · Opinion / analysis · 1 source
    Import AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space
    New benchmark shows how AI models lack intuitive visual reasoning:
17. Oct 12, 2026 · Research paper · 1 source
    Avoiding What Your Teacher Already Knows: Entity Familiarity for Deep-Search Distillation
    We first introduce Evidence-Verified Entity Familiarity (EVF), which estimates familiarity by verifying a teacher-generated factual brief against online evidence.
18. Oct 12, 2026 · Research paper · 1 source
    Embodied Skill Bench: Benchmarking How Well Agents with Embodied Skills Perform on Physical Tasks
    LLM agents increasingly control robots using VLA policies, perception models, and procedural knowledge.
19. Oct 11, 2026 · Research paper · 1 source
    Big-Big-SycoBench: A Benchmark to Detect Sycophancy Through Model Responses to Valid vs. Invalid Arguments
    We introduce Big-Big-SycoBench, which measures sycophancy through argument discernment: the ability to revise an evaluation after valid criticism while resisting invalid pressure.
20. Oct 10, 2026 · Research paper · 1 source
    Humans vs. Machines: A Comparison With AI Agents in Software Reverse Engineering
    Can an autonomous AI agent reverse engineer a binary as well as a skilled human analyst?
21. Oct 9, 2026 · Research paper · 1 source
    Humanize Olympiad Agents: A Multi-Session Workflow for Olympiad Problem Solving with Solve, Review, Wiki, Verify
    We introduce HOA (Humanize Olympiads Agent), an open, flow-level framework that coordinates multiple agent sessions through three complementary feedback channels.
22. Oct 9, 2026 · Product / feature launch · 5 sources
    [AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing
    It’s been about a year since Anthropic shipped Haiku 4.5, and with successive launches of Sonnet and Opus and Fable up to 5.5 it was seeming a little forgotten, especially as OpenAI launched Luna 6 alongside Astra and Sol 6.
23. Oct 7, 2026 · Model release · 2 sources
    [AINews] Reflection Beam - 501B-A23B American Open Model
    It’s been over a year since Reflection launched with us with big goals on coding (and hinted about their RL approach):
24. Oct 5, 2026 · Open-source release · 1 source
    Together Link: open models in the harness you already use. Start with one command today.
    Together Link brings frontier open models like GLM 5.3 and Kimi K3 into the coding agent your team already uses, cutting model spend by over 50%.
25. Oct 3, 2026 · Model release · 1 source
    [AINews] not much happened today
    AINews’ website lets you search all past issues.
26. Sep 29, 2026 · Open-source release · 1 source
    NVIDIA/TensorRT-LLM v1.3.0rc29
    Expose Nemotron-H vision-language LoRA configuration for supported inference paths #19151

Often appears with