# Qwen3.8-Max at 2.4T, Motif 3, and WeatherNext goes open > Week of 3 August 2026: Alibaba ships the 2.4T Qwen3.8-Max, Motif Technologies opens a 314B model, and DeepMind publishes cyclone forecasting in Nature. Published 13 Oct 2026. Coverage: 3 Aug 2026 to 9 Aug 2026. By Ali Asif. ## Items 1. **Qwen3.8-Max** (Alibaba, Frontier): Alibaba's flagship is a 2.4-trillion-parameter mixture-of-experts model with 95B active parameters and a 1M-token window, pitched at multi-day coding projects. [Primary source](https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B) 2. **Motif 3** (Motif Technologies, Open weights): A 314B mixture-of-experts model with 13.2B active parameters, MIT license, trained from scratch on about 12.5T tokens for agentic and terminal tasks. [Primary source](https://huggingface.co/Motif-Technologies/Motif-3) 3. **Qwen-Image-3.0** (Alibaba, Image, video, world): The standard tier of the new image series: faster and cheaper than the Pro model, with the same generation and editing interface. [Primary source](https://www.qwencloud.com/models/qwen-image-3.0) 4. **WeatherNext cyclone forecasting** (Google DeepMind, Science): A Nature paper reports three-day cyclone forecasts as accurate as older models' two-day ones, and DeepMind is open-sourcing the models. [Primary source](https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/) 5. **LFM2.5-2.6B** (Liquid AI, Open weights): A 2.69B on-device reasoning model with 131K context and 16 languages, meant for agents and data extraction on phones and laptops. [Primary source](https://www.liquid.ai/blog/lfm2-5-2-6b) 6. **Needle 2** (Cactus Compute, Agents and coding): A 45M-parameter open model in roughly 14 MB that does tool calling and structured extraction on a device. [Primary source](https://cactuscompute.com/blog/needle-2) 7. **Recursive Synthesis for Long-Horizon Terminal Tasks** (arXiv, Science): Tackles the cost of training data for terminal agents, which can run to hundreds or thousands of dollars per task. [Primary source](https://arxiv.org/abs/2608.05466) Alibaba went very big, a Korean startup went open, and DeepMind put out the kind of science result that has a use outside a lab. ## Qwen3.8-Max Alibaba released Qwen3.8-Max on 3 August. It is a 2.4-trillion-parameter mixture-of-experts model with about 95B parameters active per token, a 1M-token context and up to 128K tokens of output. Alibaba pitches it at coding and professional work, with claims about delivering complete projects that run for ten days or more and about native visual understanding across planning, execution and checking. Those are vendor claims and I'd test them on a real job before believing the ten-day figure. There is a Hugging Face page under a custom Qwen3.8-Max license, so read the terms. A related image release followed on 4 August: Qwen-Image-3.0, the faster and cheaper standard tier next to last month's Pro model. ## Motif 3 Motif Technologies released Motif 3 on 7 August. It is a 314B mixture-of-experts model with 13.2B active parameters, trained from scratch on about 12.5 trillion tokens with extra Korean, reasoning and agentic data. It has a 262,144-token native context, a built-in multi-token prediction head for faster decoding, and an MIT license. For a lab outside the biggest names, shipping from-scratch weights this size under MIT is a strong signal. ## WeatherNext, in Nature On 6 August Google DeepMind reported in Nature that its WeatherNext model predicts a tropical cyclone's path, strength and wind structure better than existing methods. Its three-day forecasts match what earlier models reached only at two days, which the company describes as roughly a decade of progress in meteorology. DeepMind is open-sourcing WeatherNext Cyclones and WeatherNext 2, plus a smaller WeatherNext 2-mini. ## Small and on-device Liquid AI released LFM2.5-2.6B on 4 August, a 2.69B reasoning model with 131,072-token context and 16 languages. Liquid does not recommend it for agentic coding, and says so plainly, which I appreciate. Cactus Compute's Needle 2 followed on 9 August: 45M parameters in about 14 MB, built for tool calls and extraction. One paper worth a look: Recursive Synthesis for Long-Horizon Terminal Tasks, which goes after the cost of building training tasks for terminal agents. ## What I'd watch next Who builds on the open WeatherNext weights first, and whether Qwen3.8-Max gets an open sibling. ## My take The weather item is the one that stuck with me. A model that gives a three-day forecast with the accuracy of an old two-day one is a real operational gain, and Saudi aviation lives on weather and wind data. On the agent side, I keep seeing the same pattern: the cost of making good training tasks is the bottleneck, which is also true of my own agents, where the hard part is writing the checks, not the prompts. ## Sources - https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B - https://huggingface.co/Motif-Technologies/Motif-3 - https://www.qwencloud.com/models/qwen-image-3.0 - https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/ - https://www.liquid.ai/blog/lfm2-5-2-6b - https://cactuscompute.com/blog/needle-2 - https://arxiv.org/abs/2608.05466 Canonical: https://heyaliasif.com/ai/2026-08-03-week-in-ai