Ai2 announces Bwen and Blama byte model checkpoints
Ai2 announces Qwen and Llama based byte model checkpoints alongside its Nature paper, with Stage 1 models for research.
AI Model Releases
New AI model releases can change the options available to developers, businesses and everyday users. ByteForward follows major launches and model updates across language, reasoning, coding and multimodal AI. This category focuses on what a release adds, how access works and which capabilities are available at launch. API availability, subscription requirements, regional restrictions and rollout stages help explain who can use a new model. Pricing, context limits and published evaluations provide further context for comparing the release with existing options. Our broader AI models coverage tracks how those systems develop after the initial announcement. Explore generative media for updates focused on AI image, video and audio creation. Browse the stories below to understand the practical details behind a model launch before deciding whether it fits your needs.
Perplexity releases two models for text and image search
Perplexity’s smaller retrieval model can query an index built with its larger counterpart.
Uratori updates a compact model for Japanese evidence checks
Uratori returns evidence judgments without generating an answer.
TwelveLabs adds first person video analysis with Pegasus 1.6
The update adds image inputs and more detailed video segmentation.
Google launches EmbeddingGemma 2 for multimodal search
The 740 million parameter release combines text, image, video and audio embeddings for local retrieval.
Google releases Nano Banana 2.1 in the Gemini API
Nano Banana 2.1 is available in the Gemini API. Google’s current documentation lists no shutdown date for the older Nano Banana 2 model.
Liquid AI adds vision to its d1 decision model
Liquid AI has enabled image inputs in its paid d1 API, expanding structured decisions to visual inspection and screenshot tasks.
Mistral Large 4 enters public preview
Mistral’s new flagship has 1.05 trillion parameters and a million token context window. Downloadable weights are planned for later in October.
Interfaze releases Lite weights for document and speech tasks
The download combines a reasoning core with specialist tools and needs a GPU with 80 GB of memory.
Reflection previews Beam with open weights planned this month
Early access is limited while Reflection completes safety evaluations.
Cactus releases Whistle for local speech recognition
The compact CPU model handles seven languages and clips up to 30 seconds.
Aleph Alpha releases Kolibri for German and English AI
Kolibri brings open weights, reasoning and tool calling, with a million token context ceiling and practical deployment limits
Fastino GLiDE brings adaptive reasoning to agent decisions
Fastino GLiDE combines adaptive reasoning with structured decisions. Its September 30 release brings new routing options and practical limits for developers.
Tavus previews Griffin for live AI video conversations
Tavus is testing Griffin Lite with selected research participants. Its video conversation results come with important limits on access and evaluation.
HeyGen Video brings scene generation and audio to its API
HeyGen Video generates complete scenes with sound from prompts and references. October launch rates depend on resolution and whether video references are used.
Perplexity releases Decider 27B weights for AI classification
Perplexity Decider 27B combines public model weights with a hosted API that returns probabilities for classification and routing tasks.
Voyage launches Rerank 3 and Lite for AI search
Voyage Rerank 3 and Lite reorder search results with updated models, familiar token prices and practical limits for long documents.
Strands Decider 2B launches for local AI agent decisions
Strands Decider 2B selects options and scores inputs for AI agent workflows, with open weights and important limits on reasoning and confidence.
Cloudflare launches Clef and Clef Flash for fast agent decisions
Cloudflare releases two open decision models for Workers AI, with typed probability outputs, visual inputs and different speed and quality tradeoffs.
Cohere launches Embed 5 Pro and Fast for enterprise search
Cohere Embed 5 pairs Pro and Fast retrieval models with a shared index, multimodal inputs and separate text and image pricing.
Decagon launches Voice 3 with its Chord speech model
Decagon pairs Voice 3 with Chord and a duplex architecture for customer calls. Here is what changed and what its early results establish.
Suno opens Speech beta for narration with generated music
Suno Speech creates spoken narration and background music in one track. The public beta is available on web and mobile, with uneven accents and timing still possible.
Microsoft launches streaming transcription and two voice models
Microsoft introduces live transcription and two voice models, with character pricing, gated cloning and important Azure public preview limits.
FLUX 3 Image launches with layout controls and 4K output
Black Forest Labs opens a dedicated FLUX 3 image endpoint with layout controls, reference editing and output up to 4K. Pricing and preservation limits matter.