New releases

Bindwell · Bindwell's foundation model for predicting experimental outcomes in molecular discovery, trained across 800M–9B sizes on about 80M experimental measurements

Cloudflare · Cloudflare's System One extraction model, which extracts, computes, estimates and forecasts values from unstructured inputs

Dream Labs · Dream Labs' World Action Model, a robot foundation model that the lab says controls unseen bimanual parallel-gripper robots zero-shot on unseen tasks

Nace.AI · Nace AI's multimodal decision model for text, images and video, available through Nace at $0.06 per 1M input tokens

GoTo (with AI Singapore) · GoTo's 27B Gemma 3 instruct model for English and Indonesian regional languages (Javanese, Sundanese, Batak Toba, Balinese), developed with AI Singapore

Inception · Inception's diffusion LLM with a 1M-token context, live on the Inception API, OpenRouter, Baseten and Vercel

Dream · Dream's open cybersecurity language models (8B and 30B-A3B), with a dedicated security mid-training stage before chat tuning

Skyfall AI · Skyfall AI's Engineering World Model for CAD and BIM tasks

Blockway · Blockway's open-weights 32B model for agentic coding, the full release after its Preview, tuned for long agent sessions

OrcaRouter · A 27B open-weight vision-capable model built for defensive cybersecurity work such as vulnerability research and incident response

NVIDIA · NVIDIA's 4B looped model: the same weights can run 1, 2 or 4 loops per token, trading compute for accuracy

Knowledgator · Knowledgator's family of instruction-following classifier encoders, from a 435M Edge model up to a 27B model

Elorian AI · Elorian AI's first frontier visual-reasoning model, built for visual thinking

Tencent · Tencent Hunyuan's dedicated long-term memory writer for AI agents, part of Hy Memory 2.0 alongside a framework and MemHub

OpenMOSS / MOSI · OpenMOSS / MOSI's text-to-image Diffusion Transformer, released as MOSS-Image-Base and MOSS-Image-SFT, with Chinese and English text rendering and controllable…

StartLux · StartLux's open-weight decision-model family: an Instant line from 0.8B to 27B and an AutoThink adaptive-reasoning line from 4B to 27B

NABLAS Inc. · NABLAS's Japanese fact-checking model (Qwen3-8B fine-tune) plus a citing agent that surfaces primary sources with the verdict; UTokyo spin-off work from GENIAC…

Datalab · Datalab's new OCR model, live on its API

IBM · IBM's 8B and 30B Granite Switch models: one checkpoint that switches through control tokens between chat, a general Decision adapter, a Guardian safety adapter…

IFM (MBZUAI) with Inception42 · A family of six open, Arabic-native models for agentic work, 0.9B to 375B, built on IFM's K2 Horizon

PrismML · PrismML's ternary on-device models in 2B, 4B and 9B sizes, based on Qwen3.5, with reasoning, vision and tool calling

NVIDIA · NVIDIA's local-agent model that fits into 24GB of memory, optimized for RTX Spark and meant to run alongside your apps

NVIDIA · NVIDIA's open multimodal model combining reasoning and tool use with text, image, document and video understanding, built for long-running AI agents

Ant Ling (InclusionAI, Ant Group) · Ant Group's open-source mixture-of-experts model from the Ant Ling team, built to turn rough briefs into pages, mobile UIs and prototypes, with up to a…