Models Editor selected
百度开源文心5.5 Flash Base，多基准测试刷新同尺寸模型 SOTA
百度于11月5日开源了其最新基座模型文心5.5 Flash Base。该模型在多个基准测试中刷新了同尺寸模型的SOTA成绩。此次开源标志着百度在AI模型领域的又一重要进展。
Why it matters: 该事件是百度最新开源模型发布，具有明确的模型名称、发布时间和性能亮点，对AI社区有直接参考价值。
百度AI开源模型大语言模型
Live intelligence
76 updatesLatest AI updates
Models
webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500
webAI has released TwIL-8B, an 8.8B dense reasoning model post-trained from IBM's Granite 4.2 8B, optimized for math, code, and instruction following. It achieves 98.2 on MATH-500 and 81.6 on LiveCodeBench, while using 38% fewer reasoning tokens than Qwen3.5-9B on AIME 2026. The model is available on Hugging Face under a non-commercial license, with GGUF builds for local CPU or modest-VRAM GPU inference.
webAIIBM Granitereasoning models
Models Editor selected
Ricursive AI Releases Ripple-65B Preview: A 65B Dense Open-Weight Model Scoring 80.5 on Terminal-Bench 2.1
Ricursive Intelligence has released Ripple-65B Preview, a 65B dense open-weight model pretrained from scratch for agentic coding. It achieves 80.5 on Terminal-Bench 2.1, the top self-reported score among US open-weight models, and is available under Apache 2.0 with three checkpoints. The model uses a dense Transformer architecture, trained on 12T tokens with up to 1M context, and claims a full R&D cost of $32M. It trails larger MoE rivals on some benchmarks like SciCode and DeepSWE.
Why it matters: This is a concrete new model release with notable performance claims and open-weight availability, offering practical value for developers and researchers in agentic coding.
AI model releaseopen-weight modelsagentic coding
Models
招商局狮子山实验室发布「驭物™ Lion-ER-35B-A3B」具身推理模型
招商局狮子山实验室发布具身推理模型「驭物™ Lion-ER-35B-A3B」，采用MoE架构，仅激活3B参数，在十项基准测试中全线提升。该模型旨在让机器人理解真实物理世界，而非依赖堆砌数据。
具身推理模型架构机器人理解
Models Editor selected
Tencent WeVisOne-VL Unifies Detection, OCR, Segmentation and Matting
Tencent's WeChat Vision team released WeVisOne-VL, an open vision language model in 2B and 4B sizes that handles detection, referring, counting, pointing, OCR, segmentation, and alpha matting in one model. It uses a 9.6M-parameter pixel decoder, eliminating the need for a separate SAM, and can output 'None' for absent objects. The model outperforms same-scale baselines on many benchmarks, including 71.3 F1 on Dense200 detection and 81.8 F1 on RefCOCOg referring, and is available under Apache 2.0.
Why it matters: This is a concrete, recent AI development with strong technical details and benchmark results, offering practical value for developers seeking a unified perception model.
Tencent WeVisOne-VLvision language modelsmulti-task perception
Models Editor selected
Axion 1.0: Open Agent Foundation Model with Simplified Training Recipe
Axion 1.0 is a fully open agent foundation model trained in just three stages: two supervised fine-tuning stages and reinforcement learning, avoiding complex techniques like DPO or on-policy distillation. It achieves competitive performance on agentic benchmarks and releases the complete training recipe, including data construction and code, for full reproducibility. The model demonstrates that strong agentic capabilities can be built with a simple, open approach.
Why it matters: Provides a concrete, recent AI development with a novel simplified training method and full open-source release, valuable for practitioners and researchers.
agentic AImodel trainingopen-source AI
Models Editor selected
Mistral Large 4: A 52B-Active Multimodal MoE for Reasoning and Agentic Tasks
Mistral Large 4 (ML4) is a new multimodal mixture-of-experts model with 52B active parameters out of 1T total, optimized for reasoning and long-horizon agentic tasks. The model is built using a scalable agent-based approach, though specific technical details are limited in the provided text. This release signals a significant advancement in large-scale MoE architectures for complex AI applications.
Why it matters: The announcement of a major new model from Mistral with a novel architecture and focus on agentic tasks is highly relevant to AI practitioners and researchers.
Mistrallarge language modelsmixture-of-experts
Models Editor selected
Deep Principle Releases Hydrogen-1.0: A 753B Open-Weight Model for Scientific Discovery
Deep Principle has released Hydrogen-1.0, a 753B-parameter open-weight scientific model built by post-training Z.ai's GLM-5.3 on domain data, laboratory records, and discovery trajectories. It reports a 38.3% score on FrontierScience-Research, outperforming GPT-6-Astra's 33.3%, and matches Kimi-K3 on GPQA-Diamond. The model is MIT-licensed and designed for chemistry, materials science, and life sciences, with deployment options on NVIDIA and AMD GPUs.
Why it matters: This is a concrete, recent release of a large open-weight model with novel training data and competitive benchmark results, offering practical value for researchers and developers in scientific AI.
AI for scienceopen-source modelsmodel benchmarking
Models Editor selected
Mistral Large 4 weights won’t get automatic guardrail updates
Mistral released the weights for Large 4 on October 29, 2026. The company clarified that these weights will not receive automatic guardrail updates, and a separate commercial license may be required for certain uses. This highlights a shift in how Mistral manages model safety updates.
Why it matters: Provides clarity on licensing and update policies for Mistral's latest model, which is valuable for developers and enterprises evaluating deployment.
Mistralopen-source AImodel licensing
Models
Anthropic to add invisible watermarks to text from older Claude models
Anthropic plans to add an invisible watermark to text generated by four of its widely used older Claude models. The change is scheduled to begin Tuesday, according to the source.
Anthropic ClaudeAI content provenancelanguage models
Models
Nace AI Open-Sources Drex Max v1: A 27B Decision Model for Local Inference
Nace AI released Drex Max v1, a dense 27B decision model, as open weights under the Nace.AI Open RAIL-M license. It returns option probabilities in a single forward pass for choice, yes/no and ordinal questions, targeting routing, classification and policy checks. The release includes bf16 weights, a Q8_0 GGUF, an affine Q4 checkpoint and forks of llama.cpp and Ollama, with the 4-bit build fitting 24 GB+ Apple Silicon Macs. Nace self-reports 67.52 on Decision Index 0.4, ahead of Drex v2.0 at 61.25.
Nace AIDrex Maxopen-weight models
Models Editor selected
Samaya Sterling Beats Claude Fable 5.1 at 11x Less
Samaya AI released Samaya Sterling, a finance research agent model with only 3B active parameters. It scores 58.9% on the company's FrontierFinance benchmark, outperforming Claude Fable 5.1 at 11x less cost. This demonstrates a highly efficient domain-specific model.
Why it matters: Provides concrete evidence of a new, efficient finance model outperforming a larger general model on a specialized benchmark.
Samaya AIfinance AIefficient models
Models Editor selected
GPT-5.6 Factorio run used 250M tokens, cost ~$52 at API prices
A tester reported that a Factorio rocket run using GPT-5.6 consumed about 250 million input tokens (95% cached) and 1.5 million output tokens. At API prices, this translates to roughly $52. The data provides a concrete example of token usage and cost for a complex task.
Why it matters: Offers real-world token and cost data for GPT-5.6, useful for developers and businesses evaluating model economics.
GPT-5.6AI model economicsFactorio
Models
Mistral Releases Weights for Trillion-Parameter Mistral Large 4
Mistral has released the weights for Mistral Large 4 (ML4), a trillion-parameter open-source AI model. Businesses and governments can download and deploy the system on their own infrastructure. The release is positioned toward cybersecurity use cases.
Mistralopen-weight modelsAI in cybersecurity
Models
Mistral releases trillion-parameter open-weight model despite dual-use warnings
Mistral has released a trillion-parameter AI model for free download, modification and local operation, placing it outside the developer's control. The developer itself warns the model could help attackers carry out real cyberattacks. The release is notable for making a very large open-weight model broadly available, together with an explicit security caveat.
Mistralopen-weight AI modelsAI cybersecurity risk
Models
NVIDIA and MIT Release Open-H3-Regenerate-2K Video Enhancement Weights on Hugging Face
NVIDIA and MIT researchers published model weights for Open-H3-Regenerate-2K on Hugging Face, a video enhancement project built on MiniMax's open-source H3 video enhancement approach. The release targets high-resolution video regeneration, with the headline claiming support for up to 4K. The supplied material contains no benchmarks, licensing details, or comparisons with existing upscalers, and the project name itself references 2K.
NVIDIAMiniMaxopen-source video models
Models Editor selected
Samaya Introduces Sterling, a 3B-Active Model for Financial Research
Samaya AI has introduced Sterling, a financial research agent model post-trained from NVIDIA’s Nemotron 3.5 Lightning, with 30 billion total parameters and 3 billion active parameters. Samaya reports a 58.9% score on its FrontierFinance benchmark at $0.61 per question, compared with 55.9% and $6.89 for Claude Fable 5.1. The technical account details model-and-harness optimization, supervised fine-tuning and reinforcement learning, but results are self-reported on Samaya’s own benchmark, and no public weights or API were announced.
Why it matters: The detailed training and orchestration recipe offers practical insights into domain-specific agent efficiency, although independent validation and public access remain absent from the supplied account.
Samaya AISamaya SterlingFrontierFinance
Models Editor selected
Edge0 Releases Apache-Licensed Audio8-TTS-Realistic for Multilingual Voice Cloning
Edge0 released Audio8-TTS-Realistic, a 601M-parameter text-to-speech model supporting 23 languages and zero-shot voice cloning from a reference clip and transcript, with weights and code under Apache 2.0. Its training uses a learned realism critic to improve speech rhythm, pauses and timing. Edge0 reports competitive multilingual error rates, but evaluation pipelines differ across models, some languages remain weak, and serving throughput and streaming have not been benchmarked.
Why it matters: The release gives developers a compact, permissively licensed multilingual voice-cloning model with deployment and fine-tuning tools, alongside important evaluation and language-quality caveats.
Edge0Audio8-TTS-RealisticVoice cloning