Microsoft Research has published and open-sourced RetroChimera, a new framework for retrosynthesis prediction that combines two complementary models to propose high-quality synthesis routes. The system integrates R-SMILES 2, a Transformer-based de-novo model, with NeuralLoc, a graph neural network (GNN) based model, using a learned ensembling strategy to rank predictions.
Google DeepMind has released AlphaGenome Atlas, a platform containing predictions for the effects of 9 billion single-nucleotide variants across the entire human genome. This resource aims to accelerate biological understanding by providing a comprehensive map of how genetic mutations affect molecular biology.
Microsoft Research has released GigaPath-Flash and GigaTIME-Flash, efficient variants of its pathology foundation models designed to enable population-scale discovery by drastically reducing computational requirements.
Zhihui AI has released GLM-5.3, an open-weights model that builds on the GLM-5.2 base architecture through post-training improvements. The update significantly enhances performance in complex coding and long-horizon tasks, establishing new state-of-the-art results on several public benchmarks.
Zai has released GLM-5.3, an open-weights model that builds on the GLM-5.2 base architecture through extensive post-training. The update delivers significant gains in complex coding and long-horizon tasks, establishing new state-of-the-art results on several public benchmarks.
ZhiPu introduces GLM-5.3-Flash, the first natively multimodal model in the GLM-5 series, featuring a hybrid architecture that combines sparse and linear attention to reduce long-context serving costs while preserving precision.
Google reports its technologies now support over 300 languages for more than 7 billion people, introducing new tools to improve speech understanding, data collection for underrepresented languages, and accessibility.
The Tencent Hy Team has open-sourced Hy4-preview, a new-generation Mixture-of-Experts flagship model comprising 770B total parameters with 49B activated per token. The architecture features 78 layers where the first uses a dense FFN and the remaining 77 use MoE with 256 routed experts, alongside a native MTP layer for speculative decoding.
DeepSeek has released V4.1-Flash, a new open-weight flagship model designed for extreme inference efficiency and low cost. The model utilizes a causal encoder-decoder architecture to significantly reduce active compute and KV/cache costs.
DeepSeek has launched DeepSeek v4.1-Flash, a new open-weight flagship model featuring a novel causal encoder-decoder architecture designed to maximize inference efficiency and reduce costs.
DeepSeek AI has released DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts model featuring a 1-million-token context window and an FP4 KV cache that reduces global cache footprint to 890 bytes per token. The model utilizes a causal encoder-decoder architecture and Compressed Sparse Attention 2 (CSA2) to significantly lower memory requirements while maintaining performance.
The Tencent Hy Team has open-sourced Hy4 preview, a new-generation Mixture-of-Experts flagship model comprising 770B total parameters with 49B activated per token. The architecture features 78 layers where the first uses a dense FFN and the remaining 77 use MoE with 256 routed experts, alongside a native MTP layer for speculative decoding.
Zhipu AI has released the GLM-5.3-Flash model, positioning it as a frontier intelligence solution with low-cost inference capabilities.
Z.ai has formally launched GLM-5.3-Flash, revealing that the previously previewed "Ox Alpha" model is its public identity. The model features 320 billion total parameters with 18 billion active, a 1 million-token context window, and native multimodal capabilities.
Z.ai has released GLM-5.3-Flash, the first natively multimodal model in the GLM-5 series, featuring a mixture-of-experts architecture with 320 billion total parameters and 18 billion active per token. The model supports image and video inputs within a 1,048,576-token context window and is available under an MIT license on Hugging Face.
Alibaba’s Qwen team has released Qwen3.8-Flash-Next, an open-weight multimodal Mixture-of-Experts model designed to preview the architecture for the upcoming Qwen4 series. The checkpoint pairs a 125B backbone with a 51B N-gram embedding table and a 4B multi-token prediction module, activating only 6B parameters per token.
Zhipu AI has released GLM-5.3-Flash, the first natively multimodal model in the GLM-5 series and the first open-weight release of the glm5_next architecture. The 320B-parameter model is trained on a 30T-token multimodal corpus and features a hybrid sparse and linear attention mechanism to reduce long-context serving costs.
Z.AI confirmed on Wednesday that the Ox Alpha model is a new iteration of its GLM series. The company stated it will release the model's weights later that night, responding to queries by Bloomberg News.
IBM has released the updated Granite 4.2 language models in 3B, 8B, and 30B parameter sizes, purpose-built for enterprise agentic workflows with native step-by-step reasoning capabilities. Alongside the language models, IBM introduced Granite Speech 5.0 Turbo CTC, a lightweight speech model optimized for high-throughput transcription on edge devices.
Xiaomi has open-sourced the MiMo-V2.6 Pro and Flash omnimodal models, which are designed to coordinate agents for constructing and visually testing interactive 3D scenes. These models can generate Blender assets, control robotic arms from camera feeds, produce frontends and presentations, assemble videos, and compose music as scores and MIDI.