embeddings

NVIDIA has released Nemotron-3-Embed-1B-BF16, a 1.14 billion parameter text embedding model supporting 34 languages with a 32,768 token context window. The model generates 2048-dimensional embeddings and was derived from Ministral-3-3B-Instruct-2512 through two rounds of structured pruning and distillation, first to 2B then to 1.14B parameters.

NVIDIA's Nemotron-3-Embed-8B-BF16 ranks #1 on the RTEB leaderboard with a 78.5% score, while the 1B variant reduces error rate by 27% over its predecessor. The open-weight models feature 32k context windows and production-ready deployment options including a Blackwell-optimized NVFP4 variant.

Vercel released version 5.0.0-canary.98 of its Google Vertex AI SDK, adding support for three new Google models: deep-research-max-preview-04-2026, deep-research-preview-04-2026, and gemini-embedding-2. The update enables developers to integrate Google's research-focused models and latest embedding model into applications using Vercel's AI SDK.

Vercel released version 4.0.0-canary.75 of its AI SDK Google package on May 30, adding support for three new Google models: gemini-embedding-2, deep-research-max-preview-04-2026, and deep-research-preview-04-2026. The update enables developers to integrate Google's latest embedding and deep research capabilities into applications built with the Vercel AI SDK.

Mistral AI released Codestral 25.08, showing 30% more accepted code completions and 10% higher retention rates. The company also shipped Devstral Small, a 24B-parameter agentic coding model scoring 53.6% on SWE-Bench Verified, alongside new embedding and IDE integration tools aimed at enterprise deployment.

IBM released two new multilingual embedding models under Apache 2.0: a 97M-parameter compact model scoring 60.3 on MTEB Multilingual Retrieval (highest in its size class) and a 311M full-size model scoring 65.2. Both support 200+ languages with enhanced retrieval for 52 languages, handle 32K-token context (64x increase over predecessors), and include code retrieval across 9 programming languages.

AWS released multimodal embedding capabilities for Amazon Nova on Bedrock, allowing manufacturing organizations to retrieve information from technical documents that combine text, engineering diagrams, and images. The model supports configurable dimensions from 256 to 3072 and processes text, images, and multi-page documents into a shared vector space.

IBM released Granite Embedding 311M Multilingual R2, a 311-million parameter dense embedding model with 32,768-token context length and support for 200+ languages. The model scores 64.0 on Multilingual MTEB Retrieval (18 tasks), an 11.8-point improvement over its predecessor, and ships with ONNX and OpenVINO models for production deployment.

Amazon Nova Multimodal Embeddings, announced October 28, 2025, now supports audio content for semantic search alongside text, images, and video. The model offers four embedding dimension options (3,072, 1,024, 384, 256) and uses Matryoshka Representation Learning to balance accuracy with storage efficiency.

Google has released Gemini Embedding 2, its first native multimodal embedding model that represents text, images, video, audio, and documents in a unified vector space. The model eliminates the need for separate embedding models across different modalities in AI pipelines.