Gemini 2.5 Deep Think
Gemini 2.5 Deep Think is an enhanced reasoning mode for Gemini 2.5 Pro, designed for complex mathematics, programming, and problems requiring creative solutions. Google announced the rollout of this…
MODEL DIRECTORY
New releases, in one place. Explore models by their publication date.
Gemini 2.5 Deep Think is an enhanced reasoning mode for Gemini 2.5 Pro, designed for complex mathematics, programming, and problems requiring creative solutions. Google announced the rollout of this…
Qwen-Flash is a text generation and reasoning model in the Qwen3 series. The official international site lists qwen-flash and the snapshot version qwen-flash-2025-07-28 in its update of August 1, 2025…
Seed Diffusion Preview is an experimental discrete diffusion language model released by ByteDance Seed. It focuses on structured code generation as a research area and is intended to test the…
Qwen3-Coder-Flash is a code generation model based on Qwen3 that inherits the coding agent capabilities of Qwen3-Coder-Plus. This API ID corresponds to the July 28, 2025 snapshot; the Alibaba Cloud…
Veo 3 Fast is Google's version of Veo 3 for faster video generation. It can generate video from text prompts or input images, and supports synchronized generation of speech and sound effects. It can…
Veo 3 is Google's generative video model. veo-3.0-generate-001 is its generally available model ID on Google Cloud Vertex AI. (docs.cloud.google.com) Supports video generation from text prompts and…
The Plus text generation and reasoning model in the Qwen3 series, using the July 28, 2025 model snapshot. (docs.modelstudio.console.alibabacloud.com) Supports thinking and non-thinking modes, which…
GLM-4.5-Air is a lightweight foundation model in the GLM-4.5 series designed for agent applications. It uses a mixture-of-experts (MoE) architecture, with 106 billion total parameters and…
GLM-4.5 is a text foundation model for agent applications. It uses a mixture-of-experts (MoE) architecture with approximately 355 billion total parameters and approximately 32 billion parameters…
Wan2.2-I2V-Plus is an image-to-video model in the Wan 2.2 series, with inference services provided through Alibaba Cloud Model Studio. It uses an input image as the first frame and can generate a…
Wan2.2-T2V-Plus is a text-to-video model in the Tongyi Wanxiang 2.2 series, with inference services provided through Alibaba Cloud Model Studio: it takes a text prompt as input and produces a video.…
Wan2.2-T2I-Flash is a fast text-to-image model in the Wanxiang 2.2 series. It takes text as input and outputs images, with inference services provided by Alibaba Cloud Bailian.…
Wan2.2-T2I-Plus is a text-to-image model in Alibaba's Tongyi Wanxiang 2.2 series. It takes text prompts as input and outputs images. According to the official description, it improves image detail,…
Tencent Hunyuan's video understanding model based on turbos-vision; Tencent Cloud TokenHub lists this version as HY-Vision-Video. (cloud.tencent.com) Supports video description and question answering…
Seed LiveInterpret 2.0 is an end-to-end real-time simultaneous interpretation model from ByteDance's Seed team. It uses a full-duplex speech understanding and generation framework and is intended for…
Qwen-Doc-Turbo is a text generation model for document data mining, designed specifically for information extraction, content moderation, classification and tagging, and summary generation. The date…
Gemini 2.5 Flash-Lite is Google's lightweight multimodal model for large-scale, low-latency applications. Google announced the official release of its stable version on July 22, 2025.…
Seed GR-3 is a vision-language-action (VLA) robotic manipulation model developed by ByteDance's Seed team for general-purpose robotics research. It uses a hybrid Transformer architecture that…
Qwen3-Coder-Plus is a code generation model based on Qwen3; this API ID points to the July 22, 2025 snapshot. The Alibaba Cloud China site announcement lists it under July 22, 2025; the international…
Qwen-MT-Plus is a flagship multilingual translation model in the Qwen series. Current official documentation states that it has been upgraded based on Qwen3, supports translation between 92 languages,…