Gemini 2.5 Flash Preview Native Audio Dialog
A preview model in the Google Gemini 2.5 Flash series, intended for real-time conversation through the Live API and capable of native audio output. It is not the same model as the separately released…
MODEL DIRECTORY
New releases, in one place. Explore models by their publication date.
A preview model in the Google Gemini 2.5 Flash series, intended for real-time conversation through the Live API and capable of native audio output. It is not the same model as the separately released…
Gemini 2.5 Flash Preview TTS is Google's preview text-to-speech model, focused on low latency, cost-effectiveness, and controllable speech generation. It takes text as input and produces audio.…
A preview text-to-speech model in the Google Gemini 2.5 Pro series, designed for high-quality, controllable speech synthesis and suited to more complex speech generation prompts. (blog.google) Accepts…
Gemini 2.5 Flash Preview is a preview model released by Google on May 20, 2025, positioned as a high-efficiency model balancing speed, cost, and reasoning capabilities. Compared with previous versions…
A large language model version in Tencent's Hunyuan TurboS series, intended for general text and reasoning tasks. This version adaptively integrates long and short chains of thought.…
codex-1 is a version of o3 optimized by OpenAI for software engineering, powering the initial research preview of Codex, a cloud-based coding agent. It is trained with reinforcement learning on…
hunyuan-t1-vision is Tencent Hunyuan's multimodal understanding deep-thinking model. Tencent Cloud's product updates record that its official release was on May 16, 2025. (cloud.tencent.com) The model…
Qianfan-Sug is a dialogue prediction model developed in-house by Baidu's Qianfan platform, specialized in recommending the next question in multi-turn conversations. Baidu lists it as a text…
Wan2.1-VACE-Plus is a unified video creation and editing model in Wanxiang 2.1. It accepts conditional inputs such as text, images, and video, and outputs video. The official announcement lists it as…
The image-to-video version of Seedance 1.0 lite is a video generation model with a small parameter count developed by ByteDance and optimized for efficiency and cost. This API ID corresponds to image-…
The text-to-video version of Seedance 1.0 lite is a video generation model with a small parameter count, optimized for generation speed and cost. It takes a text prompt and optional parameters as…
The May 8, 2025 snapshot of Qwen's real-time multimodal understanding and generation model, designed for real-time audio interaction. (docs.modelstudio.console.alibabacloud.com) Supports text, image,…
Gemini 2.0 Flash Image Generation preview model, released by Google on May 7, 2025, for developers building conversational image generation and editing applications. At release, it was available…
A Gemini 2.5 Pro preview update released early by Google on May 6, 2025, focused on improving coding capabilities, particularly for interactive web applications and front-end interface development.…
kimi-thinking-preview is a preview version of a long-thinking model released by Moonshot AI. The official description presents it as a model with multimodal and general reasoning capabilities that…
DeepSeek-Prover-V2-7B is a 7B-parameter model for formal theorem proving in Lean 4, built on DeepSeek-Prover-V1.5-Base, and supports a context length of up to 32K tokens. The model weights are…
DeepSeek-Prover-V2-671B is a mathematical reasoning model for formal theorem proving in Lean 4. It was trained on DeepSeek-V3-Base and uses the same architecture as DeepSeek-V3. Official model weights…
aitryon-plus is a virtual try-on image generation model from Alibaba Cloud Bailian, released on April 28, 2025. (docs.bailian.console.aliyun.com) The model takes flat-lay images of clothing and front-…
An introduction in this language is coming soon.
A visual understanding and OCR model trained on Qwen-VL, using a unified model to handle image and text recognition, parsing, and processing tasks. This version is a snapshot from April 13, 2025.…