Meituan made LongCat 2.5 available on the LongCat API after a September preview. The 1.6-trillion-parameter mixture-of-experts model activates about 48 billion parameters per token and combines image understanding with long-horizon coding and tool use across several agent harnesses.
Why it matters: The material change is availability: teams can now test the final model through documented chat, responses and Anthropic-compatible interfaces instead of relying on a preview announcement.
Liquid AI released six roughly 2B-class LFM3 models for text, vision, audio, omni and calibrated decision tasks. Open weights, API access and day-one support span Hugging Face, llama.cpp, MLX, vLLM and multiple hardware vendors.
Why it matters: A shared compact backbone across modalities creates a practical local-specialization option: teams can compare one small family for on-device agents, video, speech and decision outputs instead of stitching together unrelated large models.
DeepSeek moved the deepseek-flash API alias plus its web and app experiences to DeepSeek-V4.1-Flash-1017 without changing the API call or price. The vendor reports shorter runtime and output plus higher agent, coding, reasoning, automation and visual-tool benchmark scores than the prior V4.1 Flash build.
Why it matters: A mutable production alias changed underneath existing integrations. Even with the same endpoint and price, output length, latency, tool behavior and task quality can move, so cached evaluations and operating budgets may no longer describe production.
Meta released BF16 and NVFP4 weights for Muse Spark Open, a coding-oriented multimodal mixture-of-experts model with roughly 816B total and 42B active parameters, text, image, audio and video inputs, text output and up to one million tokens of shared context; AWS added both variants to SageMaker JumpStart.
Why it matters: Open weights and multiple runtime paths make a very large multimodal coding model inspectable and deployable outside one hosted API. Its scale, license and output of reasoning traces make hardware cost, data handling and trace exposure part of the adoption…
Mistral completed the promised Mistral Large 4 weight release with official FP8 and NVFP4 checkpoints plus community GGUF conversions. The 1-trillion-parameter, 49-billion-active multimodal MoE keeps a 1-million-token context window and supports function calling and configurable reasoning levels.
Why it matters: The 29 October release turns the earlier API preview into an inspectable self-hosting option across several runtimes and precision levels. Operators can now test exact artifacts and licenses, but the model remains hardware-intensive and hosted commercial use…
Xiaomi released native multimodal MiMo-V2.6-Pro and Flash through its API, desktop client and open weights, alongside a 9B distill checkpoint, technical report, more than 7,000 reinforcement-learning task environments, an end-to-end training framework and composable agent harnesses. The models accept text, image, video and audio with one-million-token context.
Why it matters: The release exposes both large multimodal checkpoints and parts of the training machinery behind long-horizon coding, web, knowledge and cybersecurity agents. That makes the work more reproducible than an API-only launch, while the trillion-parameter Pro…
xAI's Grok 4.7 remains available through its own API, and AWS added it to Amazon Bedrock in AWS GovCloud (US) on 19 October. The xAI route offers a 500K-token context and US regional endpoint; the new Bedrock route gives US public-sector teams supported GovCloud console and API access.
Why it matters: GovCloud distribution changes the procurement, residency and control path for regulated US workloads without independently changing the model's underlying capability evidence.
Researchers inserted Android-style permission dialogs into real GUI tasks and found that requester identity and task framing sharply changed grants; prompt instructions reduced unnecessary permissions inconsistently and sometimes blocked legitimate ones.
Why it matters: A task-completion model has a built-in incentive to click Allow even when the requested capability is not necessary or appropriately scoped.
Thinking Machines released a 975B-parameter mixture-of-experts model with 41B active parameters, one-million-token context and native text, image and audio input. Apache-2.0 BF16 and NVFP4 checkpoints support Transformers, vLLM, SGLang and llama.cpp paths.
Why it matters: A multimodal open-weight model with several day-one runtimes gives teams a controllable foundation for coding, retrieval and media workflows that cross input types.
Google released Gemini 3.5 Flash across the Gemini app, Search AI Mode, Antigravity, the Gemini API and enterprise products. Google reports that it beats Gemini 3.1 Pro on several coding, agent and multimodal benchmarks while producing output four times faster than other frontier models.
Why it matters: A fast model capable of long-horizon tool use changes routing economics: many workflows no longer need the largest model at every step.
Moonshot AI released Kimi K2.6 for long-horizon coding, coding-driven design, proactive agents and expanded agent swarms. It was available through Kimi, the app, API and Kimi Code, with open weights; Moonshot described swarms of up to 300 sub-agents and 4,000 coordinated steps.
Why it matters: The release shifts the open-model frontier from answering coding questions toward maintaining state across hours, using tools and coordinating heterogeneous workers. It also gives builders a downloadable alternative for persistent agent systems.
Google released Gemma 4 in Effective 2B, Effective 4B, 26B mixture-of-experts and 31B dense variants under Apache 2.0. The family targeted reasoning, multimodal and agentic workflows across mobile devices, laptop GPUs, workstations and accelerators.
Why it matters: A consistent, permissively licensed model family lets developers choose the smallest deployable model for privacy, latency and cost. Edge and workstation deployment also creates a credible fallback when hosted APIs are unavailable or inappropriate.
Google introduced an upgraded Gemini Deep Think mode for difficult science and engineering problems, then released Gemini 3.1 Pro in preview across developer, cloud and consumer surfaces. Deep Think targeted Google AI Ultra subscribers and select API users, while 3.1 Pro became the broadly testable foundation for complex reasoning and agent workflows.
Why it matters: Reasoning is becoming a configurable service tier. Teams can use a general model for most work and reserve expensive multi-path inference for problems where extra search and verification improve the answer.
Moonshot AI released Kimi K2.5 as an open-weight native multimodal model for coding, visual debugging, office artifacts and tool use. Its research-preview Agent Swarm can dynamically create up to 100 sub-agents and coordinate as many as 1,500 tool calls; the model was available through Kimi, its API, Kimi Code and downloadable weights.
Why it matters: This was a concrete attempt to train orchestration rather than hand-code a fixed collection of agents. It also made visual-to-code and parallel research capabilities available outside a closed Western frontier API.
Mistral released Mistral Large 3, a 675B-parameter mixture-of-experts model with 41B active parameters, plus 14B, 8B, and 3B Ministral 3 models, all under Apache 2.0.
Why it matters: One permissively licensed family covered frontier-scale serving, enterprise customization, laptops, and edge devices, reducing the need to adopt unrelated model stacks for each tier.
Google made Gemini 2.0 Flash generally available through the Gemini API and Vertex AI, while previewing 2.0 Pro Experimental, Flash-Lite, and a thinking variant. The family emphasized multimodality, long context, tool use, and multiple price/performance tiers.
Why it matters: A model portfolio let builders route extraction, interactive work, and harder reasoning to different endpoints under one platform.