GLM-5.3-FlashX hits 200 tokens/sec on B.AI via 320B sparse model with 1M context and native multimodal support.

B.AI rolls out GLM-5.3-Flash with 320B parameters and 1M context plus five other free models for developers at chat.b.ai.

According to Richard Seroter, the new Google AI Gemini Embedding 2 model introduces the ability to process interleaved modalities in a single request, allowing users to obtain embeddings for both images and their corresponding text captions simultaneously. This advancement could significantly enhance AI model efficiency and multimodal application development.

According to @DeepLearningAI, Amazon introduced the Nova 2 family — Pro, Omni, Lite, and Sonic — delivering competitive multimodal reasoning and generation capabilities, which the source describes as rivaling leading systems in functionality (source: @DeepLearningAI). Nova Forge enables customers to combine their own data with Amazon checkpoints for custom training, indicating first‑party support for tailored enterprise model fine‑tuning within Amazon’s stack (source: @DeepLearningAI). Nova Act adds browser‑automation agents that can navigate websites, fill forms, and extract data, expanding enterprise‑grade agent workflows highlighted by the source (source: @DeepLearningAI). Early benchmarks cited by the source show Nova 2 Pro rivaling leading models on several tests, underscoring competitive parity claims relevant to model selection decisions (source: @DeepLearningAI). The source makes no mention of any cryptocurrencies or tokens in connection with the announcement, while centering on AI agents and multimodal model capabilities that traders often monitor for sentiment in AI‑exposed markets (source: @DeepLearningAI).

According to @AIatMeta, Meta is open-sourcing the Perception Encoder Audiovisual (PE-AV), the technical engine that powers SAM Audio’s state-of-the-art audio separation capabilities. Source: AI at Meta, X, Dec 18, 2025. The post specifies PE-AV builds on the earlier Perception Encoder and integrates audio with visual perception to achieve audiovisual source separation, highlighting a multimodal approach relevant to real-world media processing. Source: AI at Meta, X, Dec 18, 2025. The announcement does not disclose any cryptocurrency, blockchain integrations, token partnerships, repository link, or license details, indicating no direct on-chain tie-in or immediate code access information within the post. Source: AI at Meta, X, Dec 18, 2025.

According to @demishassabis, Gemini 3 Pro is a state of the art vision and multimodal AI model with strong document, screen, image, video, and spatial understanding and is available now in the Gemini App. source: Demis Hassabis on X Dec 7 2025 Gemini is developed by Google DeepMind within Alphabet, so equity traders may track Alphabet ticker GOOGL for AI narrative exposure tied to the Gemini product line. source: Google blog Introducing Gemini Dec 2023; Alphabet 2023 Form 10 K Crypto traders can add AI linked tokens FET RNDR and AKT to their watchlist given their direct ties to AI agents GPU rendering and decentralized compute as disclosed by their official documentation. source: Fetch ai documentation; Render Network documentation; Akash Network documentation

According to @karpathy, Gemini Nano Banana Pro solved chemistry exam questions directly from an image of the exam page, correctly parsing doodles and diagrams, with ChatGPT later judging the answers correct except for a nomenclature fix on Se2P2 and a spelling correction for thiocyanic acid, source: Andrej Karpathy on X, Nov 23, 2025. The demo evidences in-image multimodal parsing and reasoning on dense document layouts, which aligns with Google’s Gemini family positioning and the inclusion of Nano in the product lineup, source: Andrej Karpathy on X, Nov 23, 2025; Google DeepMind Gemini introduction, Dec 2023. Historically, prominent AI capability reveals have coincided with rotations into AI-linked crypto assets such as RNDR and FET and related equities after major AI news, source: Reuters reporting on AI token rallies during the ChatGPT surge in Feb 2023 and after Nvidia earnings in May 2024. Traders may watch Alphabet GOOGL and AI infrastructure tokens for narrative momentum if this demo draws broader attention, while noting the accuracy risk highlighted by the Se2P2 naming and spelling errors, source: Andrej Karpathy on X, Nov 23, 2025; Reuters Feb 2023 and May 2024.