$ articles --topic=model-releases

Model Releases

49 AI articles about Model Releases.

Speed Up Image Generation with FLUX 3 Image Klein on NVIDIA Blackwell GPUsLearn how to reduce diffusion model inference latency when running FLUX Image Klein on NVIDIA Blackwell GPUs. The piece covers why sequential passes and high-resolution attention slow generation in design and image editing tools, and how to speed up the pipeline.Model Releases · via NVIDIA Technical Blog · 1d ago1 min [AINews] Inkling 2 Small: 17A-594B trained in 68 hours on 12T tokens, beating Muse SparkThis roundup covers Thinking Machines' Inkling 2 Small, a 594B-total, 17B-active multimodal MoE, including its benchmark gains over the predecessor, Tinker pricing, and the promised Apache 2.0 weights release. It also explains how the model compares with peers on coding, arenas, and domain evals, and the alignment problems early checkpoints showed.Evals · via Latent Space · 1d ago7 min Introducing the New Auto and Auto Code RoutersExplains the new Auto Router, rebuilt on a fresh technical foundation, and Auto Code, a router tuned for coding tasks. Both choose a model and reasoning effort for each request by weighing eleven factors, including live pricing, provider health, benchmarks, and cache state, within the models your governance policies permit.AI Coding Tools · via OpenRouter · 1d ago1 min [AINews] Behind Latent Space: Kill My SaaS, Jev & AI in 2026 | AIEi ShanghaiThis issue of AINews covers claims that frontier LLMs can read and write neural network weights directly, including a Qwen edit that lifted MATH-500 scores and its skeptical critiques. It also surveys new model releases, independent evals, provable inference from Anthropic, and audits that found flaws in popular benchmarks.Evals · via Latent Space · 3d ago7 min vLLM Delivers Day-0 Support for Kimi K3.1 ProThis article explains how vLLM added day-0 support for Kimi K3.1 Pro. It covers the model's sparse attention on a KDA hybrid, its packed KV cache using FP8 or NVFP4 latents and MXFP4 index keys, token-sharded prefill, decode context parallelism, and fused low-latency decode.Model Releases · via vLLM · 3d ago1 min Introducing Images 2.6Images 2.6 is a new release that speeds up image generation, adds GIF creation, and offers more precise editing. It also introduces new API controls that work across ChatGPT, Codex, and the API.Model Releases · via OpenAI News · 3d ago1 min llm 0.37.1This release of LLM, the command-line tool and Python library for working with large language models, is a bugfix update. The notes credit contributions from Steven Huang, Tao An, Arjan Mossel, and XQ. The entry gives no detail on which bugs were fixed.Model Releases · via Simon Willison · 4d ago1 min Which MiniMax H3 model should you use?Compares the three MiniMax H3 models, H3 Highspeed, H3 Lite, and H3, and their different prices. It explains which situations suit each one so you can pick the right model for your needs.Model Releases · via Replicate · 4d ago1 min Open-Weight vs Closed-Weight Models: What's the Difference?This piece explains how open-weight models, whose trained weights anyone can download and run, differ from closed-weight models that only their maker provides. Using a year of OpenRouter usage data, price catalogs and benchmark scores, it shows how that difference affects cost, the choice of providers and output quality.Model Releases · via OpenRouter · 7d ago1 min Pydantic Validation: Next StepsCovers the plans for Pydantic V3, including which breaking changes it will allow (correctness fixes and easily reversible opt-ins) and the removal of V1 deprecations and the pydantic.v1 namespace. It also explains that V1 support has ended, how V2 will get fixes after V3 ships, and what enterprise support offers.Model Releases · via Pydantic · 8d ago3 min Weaviate 1.40 ReleaseWeaviate v1.40 makes named vector dropping and 4-bit Rotational Quantization generally available, adds multi-vector MUVERA support on HFresh and a new PathSeer filter strategy for HNSW. It also flags a breaking change to vector configuration defaults to check before upgrading.Model Releases · via Weaviate · 8d ago8 min Le Chonk: Built for Coders, Trusted by Defenders.The article announces that the weights of Mistral Large 4, nicknamed Le Chonk, are now publicly available. It describes the model as a 1-trillion-parameter multimodal system, presented as the strongest open-weight model for coding and trained in Europe.AI Coding Tools · via Mistral · 9d ago1 min