English text is being prepared.

Reflection's Open-Weight Model Beam Launches on Ollama, Free for a Limited Time

Reflection's first open-weight model, Beam, is officially released, offered free for a limited time with Ollama as a day-zero partner and with zero data retention (ZDR). Beam is an MoE model with approximately 501 billion total parameters, about 23 billion activated per token, and a 1 million token context window. It targets software engineering, reasoning, and long-horizon agentic tasks, and is released under the Apache 2.0 license. Users can run beam:cloud via Ollama Cloud and connect it to tools such as Claude Code, Codex, and OpenCode.

Meta open-sources 816B-parameter model Muse Spark Open, adapted for Apple MLX to run on Mac

Meta has open-sourced the 816B-parameter Muse Spark Open on Hugging Face, sharing the same base model as Muse Spark 1.3 with added safety post-training. The model has 256 expert networks per layer, routing each token to 8 of them, so deployment requires 816B-parameter-level VRAM but runs at the speed of a 42B model; the 4-bit NVFP4 version requires at least 700GB of GPU VRAM, and Apple has adapted it for its MLX framework, with two M5 Ultra Mac Studios each providing up to 512GB of unified memory and 1.2TB/s bandwidth to carry it. It tops the MCP-Atlas tool-calling test at 89.2, and the license allows free commercial use but requires attribution as Built with Muse Spark Open; API model sellers with over $20 million in revenue over 12 months must sign a separate agreement with Meta.

English translation is being prepared.

Swiss AI releases the Apertus 1.5 multimodal model, expanding context to 262k tokens

Swiss AI has released Apertus 1.5, which builds on 1.0 by adding native image and audio understanding, a thinking mode, and a 262,144-token context window—4 times larger than 1.0. The model uses early fusion to convert images and audio into discrete tokens and feed them into the same sequence as text; the architecture is unchanged, and it only understands multimodal input while outputting text only. The team says the 8B version outperforms open-source models of the same scale on instruction following and multilingual tasks, its multimodal performance is close to Gemma 4 12B, and it has also released the training data mixture, preference data, and intermediate checkpoints from each stage.

English translation is being prepared.

Google Gemini 4 Argon is fully open to developers worldwide

On October 13, US local time, Google Gemini 4 Argon opened to developers and enterprises worldwide, callable via platforms such as the Gemini API, Google AI Studio, and Google Antigravity; Gemini Pro and Ultra subscribers can also experience it in the Gemini app and AI Mode in Search. Google also disclosed several application cases, including using Argon to refactor a Chromium legacy code directory of about 6 million lines, raising Gemma 4 E2B inference throughput on a single A100 from about 37,900 tokens per second to over 108,000, and building a Real2Sim simulation environment from real robot teleoperation videos. Limited-time API pricing is 2 USD per million input tokens and 10 USD for output, rising to 4 USD and 20 USD respectively from December 1; the model offers a unified 1 million token context window.

Ai2 Releases MolmoMotion, a Language-Guided 3D Motion Prediction Model

Ai2 releases MolmoMotion: by inputting video frames, 3D points on objects, and action descriptions, it can predict the trajectories of these points in 3D space over the next few seconds, outperforming existing prediction methods. It also open-sources the MolmoMotion-1M dataset (derived from 1.16M videos) and the PointMotionBench benchmark (2.7K video clips), and releases the model weights. In simulation, a control policy based on MolmoMotion achieved a 76.3% success rate on pick-and-place tasks, higher than the 56.0% achieved by one based on Molmo 2.

Anthropic releases Claude Sonnet 5.5, 30% faster than Sonnet 5 with unchanged per-token pricing

Anthropic has released Claude Sonnet 5.5, the second model in the Claude 5.5 family. It is smarter and more efficient than Sonnet 5 and 30% faster, with unchanged per-token pricing; because it needs fewer tokens to complete the same work, costs for most tasks drop by up to 30%. The official recommendation is to use Sonnet 5.5 for well-scoped everyday tasks such as bug fixing and rapid feature iteration, while complex long-running work should still use Opus 5.5; Haiku 5.5 will join the family in the coming weeks. The model ID is claude-sonnet-5-5, and it is already available on platforms including the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry. Starting with v2.1.284, the sonnet alias in Claude Code points to this model, which runs at medium effort by default and natively supports a 1M context window.

Ant Ling Open-Sources Native Multimodal Model Ling-3.0-flash-VL

Ant Ling has released and open-sourced Ling-3.0-flash-VL, the first native multimodal large model in the Ling series. Built as an extension of the Ling-3.0-flash MoE architecture, it has 124B total parameters and 5.5B activated per inference, natively supports image, text, and video inputs, and has a 256K Token context window. The company says native multimodal joint training improved the model's score on the Artificial Analysis Intelligence Index v4.1.1 by 4 points over the text-only version, and it introduces a visual feedback loop of observe, act, verify, and correct for front-end replication, GUI Agent, and medical report interpretation. The model is now available for free trial on Ling Studio, and the BF16 and FP8 versions have been open-sourced on Hugging Face and ModelScope.

GPT-6 Family's First Model Astra Released, Author's Hands-On Test Consumed 4 Billion Tokens

OpenAI released Astra, which the author calls the beginning of the GPT-6 family; it performs outstandingly on ARC-AGI-3 and scored highest on Zapier's AutomationBench, with pricing the same as Fable 5.1. During weekend trials, the author consumed 4 billion tokens and found its capability fluctuates noticeably—it is a showcase model rather than one that reliably does work. Astra in Codex can now continue advancing tasks without waiting for an answer, and OpenAI also claims it has achieved its automated research intern goal, with the next goal being an automated AI researcher by March 2028.

Meta's Muse Glimmer 30B Launches on Fireworks, Targeting Persistent Agents

Fireworks has launched Meta's Muse Glimmer 30B dense model under the Apache 2.0 license, supporting 128K+ token context and native image understanding, targeting persistent agents that require multi-turn tool calling. The model scores 75.5 on MCP Atlas, 74.6 on DeepSearch QA, and 51.2 on SWE-Bench Pro, all higher than the corresponding results of Gemma 4 31B and Qwen 3.6 27B; these benchmark figures are reported by Meta. Fireworks also provides recommended settings of temperature 1.0, top_p 0.95, and top_k 64, and supports adjusting reasoning intensity in the system prompt.