Google’s new frontier model leads an independent benchmark, but its initial rollout is restricted to selected cyber defenders.
Mistral's October 29 release makes Large 4 weights downloadable. Coding results, infrastructure requirements and a model-as-a-service license restriction shape the trade-offs.
Giotto's October 27 launch offers European cloud and private-deployment options. Its published benchmark scores and commercial terms need separate scrutiny.
NVIDIA's October 28 speech-model release adds control over transcript format, with open weights and important deployment and language-support limits.
OpenBMB's October 28 OCR suite pairs a compact document parser with a correctness checker. Open weights are available, but review thresholds still need testing.
Alibaba's October 27 Flash release expands context capacity and adds a useful voice-agent tool combination, while retaining audio-history limits.
PFN's October 27 model upgrade adds research tools and deployment choices. Availability, vendor claims and free-tier data use deserve separate checks.
Tencent's October 26 release brings open translation weights and serving tools. Speech synthesis, benchmark interpretation and deployment support need separate checks.
Nanbeige's October 24 open-weight release offers another coding-agent option. Download size, vendor benchmarks and the smaller-model roadmap need separate scrutiny.
Microsoft's October 27 MAI Image 2.7 preview improves image-editing results, while text-to-image access and production API availability remain separate questions.
Baidu's October 27 model release targets cross-page document parsing with open weights and a hosted beta. Deployment trade-offs and service limits still matter.
Liquid AI's October 27 LFM3 release brings downloadable compact models for text, vision, audio and decisions, with important limits for local deployment.
Anthropic's October 26 S-1 reveals first-half revenue, financing-related accounting charges and proposed voting rights. Offering terms remain undecided.
Google DeepMind's October 27 Genie 4 announcement adds audio and interactive dialogue, but access remains limited to trusted testers.
OpenAI's October 26 release adds a lower-priced visual and computer-use model. Here is what is available, what it costs and what still needs testing.
Receive the latest AI news
