All services

Motion control: the character in your image performs the movements and expressions of your video (3–30 seconds).

Makes the person in your image perform the motion of a reference video, with its audio.

Swaps the person in a video for the one in your image, keeping motion, camera and sound.

Full songs with vocals, or instrumentals. Returns two takes. Attach a text output as lyrics, and the prompt becomes the style.

Speech to text for audio or video, in any of 50+ languages.

Captions, alt text, OCR or answers about an image.

Small, quick and cheap. Good for rewrites, titles and short copy.

General writing.

Aion 3.0 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models.

Aion 3.0 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the DeepSeek family of models.

Aion 3.5, the successor to Aion 3.0, is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models, with a longer context window.

Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models.

Seed 2.1 Turbo (Dola-Seed-2.1) is ByteDance’s next-generation multimodal model for the coding and agent era, with engineering-grade code delivery, long-horizon agent execution, and

DeepSeek-V3.2 is an efficient large language model with DeepSeek Sparse Attention (DSA) for long contexts.

DeepSeek V4 Flash is an efficiency-optimized 284B-parameter Mixture-of-Experts model with 13B active parameters and a 1M-token context window.

DeepSeek V4 Flash 0731 follows the 0423 release of DeepSeek V4 Flash, an efficiency-optimized 284B-parameter Mixture-of-Experts model with 13B active parameters and a 1M-token context window.

DeepSeek V4 Flash 0731 follows the 0423 release of DeepSeek V4 Flash, an efficiency-optimized 284B-parameter Mixture-of-Experts model with 13B active parameters and a 1M-token context window.

DeepSeek V4 Pro is a 1.6T-parameter Mixture-of-Experts model with 49B active parameters and a 1M-token context window.

DeepSeek V4 Pro is a 1.6T-parameter Mixture-of-Experts model with 49B active parameters and a 1M-token context window.

DeepSeek V4.1 Flash is a multimodal Mixture-of-Experts model with 552B backbone parameters (8B active on input, 16B on output) and a 1M-token context window.

DeepSeek V4.1 Flash is a multimodal Mixture-of-Experts model with 552B backbone parameters (8B active on input, 16B on output) and a 1M-token context window.

DeepSeek V4.1 Flash is a multimodal Mixture-of-Experts model with 552B backbone parameters (8B active on input, 16B on output) and a 1M-token context window.

Llama 3.2 3B is a lightweight model from Meta's Llama 3.2 family, sized for fast, low-cost tasks such as summarization, classification and rewriting.

Llama 3.3 70B is Meta's instruction-tuned 70-billion-parameter Llama model, which Meta reports performs close to its much larger Llama 3.1 405B on many tasks.

Muse Spark Open is an open-weight reasoning model from Meta for coding and agentic workflows.

Muse Wonder is a frontier reasoning model from Meta for complex personal and professional agentic tasks.

MiniMax-M2.5 is a large language model optimized for coding, agentic workflows, and modern application development with enhanced reasoning capabilities.

MiniMax-M2.7 is a large language model designed for autonomous, real-world productivity with advanced agentic capabilities through multi-agent collaboration.

MiniMax-M3 preview is a 1.4T-parameter frontier model from MiniMax for coding, agentic workflows, and complex reasoning, served at fp8 with a 512K context window.

MiniMax M3.1 Flash is a frontier model for coding, agentic workflows, and complex reasoning, with a 1M-token context window, text, image, and video input, and tunable thinking dept

Mistral Small 3.2 is a 24B parameter model optimized for efficiency and performance.

Kimi K2.5 is an open reasoning model from Moonshot AI, featuring trillion-parameter Mixture-of-Experts architecture with 32B active parameters and a 256K-token context window.

Kimi K2.6 is an open-source, native multimodal agentic model from Moonshot AI with 1T total parameters and 32B active parameters.

Kimi K2.7 Code is Moonshot AI's coding-focused agentic model built on Kimi K2.6, with 1T total parameters and 32B active parameters.

Kimi K3 is an ultra-large-scale, open-weight multimodal reasoning model from Moonshot AI.

Kimi K3 is an ultra-large-scale, open-weight multimodal reasoning model from Moonshot AI.

Kimi K3.1 Pro is Moonshot AI's flagship multimodal reasoning model and the post-trained successor to Kimi K3, with a 1M-token context window.

Kimi K3.1 Pro Highspeed is the faster variant of Kimi K3.1 Pro, Moonshot AI's flagship multimodal reasoning model.

NVIDIA Nemotron 3 Nano 30B is a compact and efficient language model from NVIDIA, optimized for fast inference while maintaining strong performance across diverse tasks.

NVIDIA Nemotron 3 Ultra is built for frontier reasoning, orchestration, coding agents, deep research, and complex enterprise workflows.

NVIDIA Nemotron 3.5 Super VL is an open multimodal vision-language model for specialized agents.

Answers directly, without a reasoning step.

Reasons step by step before it answers.

Turbo variant of Qwen3 Coder 480B, optimized for faster inference on code tasks.

Optimized for speed and efficiency.

Qwen 3.5 35B A3B is a highly efficient MoE model with 35B total parameters and only 3B active parameters.

Qwen 3.5 is a reasoning model from Alibaba featuring a 397B parameter Mixture-of-Experts architecture with 17B active parameters.

The Qwen 3.6 27B native vision-language dense model builds upon the 3.5-27B version, with key improvements in agentic coding capabilities and enhanced STEM reasoning and inference skills.

Qwen 3.6 35B A3B is a fast mixture-of-experts model with 35B total parameters and ~3B active per token.

Qwen 3.7 Max is the largest model in the Qwen 3.7 series, with deep thinking, function calling, prompt caching, and multimodal input support for images and video.

Qwen 3.7 Plus, the successor to Qwen 3.6 Plus, is a reasoning model from Alibaba with exceptional performance across coding, reasoning, and general knowledge tasks.

Qwen 3.8 2.4T is Alibaba's open-weight 2.4-trillion-parameter MoE model (95B active), with major gains in software engineering, research, and long-horizon agentic tasks.

Qwen 3.8 27B is a native vision-language dense model with 27B parameters.

Qwen 3.8 Flash is a multimodal model in the Qwen family, pairing strong reasoning and generation with remarkable speed.

Qwen 3.8 Max is Alibaba's flagship 2.4-trillion-parameter MoE model, with major gains over Qwen 3.7 Max in software engineering and office-productivity workflows and strong long-ho

Qwen3-VL 235B vision-language model with MoE architecture.

Hy4 Preview is a 770B-parameter Mixture-of-Experts flagship model from the Tencent Hy team, with 49B parameters active per token and a 1M-token context window.

Pareto is a blended model that runs several frontier models in parallel on every request and returns one answer.

Grok 4.20 is xAI's multimodal reasoning model with strong tool use, structured output support, and a 2M-token context window.

Grok 4.20 Multi-Agent is a variant of xAI Grok 4.20 designed for collaborative, agent-based workflows.

Grok 4.3 is a reasoning model from xAI with function calling, structured outputs, and a 1M-token context window.

Grok 4.5 is xAI's intelligent coding model for agentic software engineering and workflow tasks, with function calling, structured outputs, and a 500K-token context window.

Grok 4.6 is xAI's multimodal chat and reasoning model with function calling, structured outputs, adjustable reasoning effort (low/medium/high/xhigh), and a 500K-token context window.

Grok 4.7, the successor to Grok 4.6, is xAI's multimodal chat and reasoning model with function calling, structured outputs, adjustable reasoning effort (low/medium/high/xhigh), an

xAI's fast coding model trained specifically for agentic coding, currently in early access.

MiMo-V2.5 is Xiaomi's native omnimodal model with strong agentic capabilities, supporting text, image, video, and audio understanding in a unified architecture.

MiMo-V2.6-Flash is Xiaomi's efficiency-balanced omnimodal model: a sparse Mixture-of-Experts with 309B total and 15B activated parameters, natively understanding text, image, video

GLM-4.6 is a large language model developed by Z.ai, featuring strong reasoning capabilities and support for multiple languages.

GLM-4.7, the successor to GLM-4.6, is a large language model developed by Z.ai, featuring strong reasoning capabilities and support for multiple languages.

GLM-5, the successor to GLM-4.7, is a large language model developed by Z.ai, featuring significantly enhanced reasoning capabilities, improved instruction following, and support f

GLM-5 Turbo is a fast inference model from Z.ai tuned for strong performance in agent-driven environments and production coding workflows.

GLM-5.1, the successor to GLM-5, is a large language model developed by Z.ai, featuring significantly enhanced reasoning capabilities, improved instruction following, and support f

GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.

GLM-5.3 Flash is a reasoning model designed for coding, sustained agentic work, and production workloads.

GLM-5V-Turbo is Z.ai's first native multimodal agent foundation model, built for vision-based coding and agent-driven tasks with image and text inputs.

Orpheus TTS is Canopy Labs' open-weight speech model, built on Llama 3B for natural, emotive delivery.

ElevenLabs Turbo v2.5 is ElevenLabs' low-latency multilingual speech model, balancing quality and speed for real-time use.

Gradium TTS is Gradium's text-to-speech model for natural, low-latency voice generation.

Kokoro is hexgrad's open-weight 82M-parameter text-to-speech model.

Inworld TTS-1.5 Max is the higher-quality tier of Inworld's TTS-1.5 speech model, made for expressive delivery in games, companions and voice agents.

Clone your voice from a short recording and generate natural speech in it across 30+ languages.

Qwen 3 TTS 0.6B is the smaller of Alibaba's two open-weight Qwen3-TTS models, built for fast, low-cost speech.

Qwen 3 TTS 1.7B is the larger of Alibaba's two open-weight Qwen3-TTS models, trading some speed for higher-quality speech.

Chatterbox HD is Resemble AI's open-weight text-to-speech model, known for expressive, natural delivery.

xAI TTS v1 is xAI's text-to-speech model, made for narration, assistants and voice apps.

ElevenLabs Scribe V2 is ElevenLabs' speech-to-text model, built for accurate transcription across many languages.

Parakeet TDT 0.6B v3 is NVIDIA's open-weight speech recognition model, with 600 million parameters and support for 25 European languages, which it detects on its own.

xAI Speech to Text v1 is xAI's speech recognition model, which transcribes recorded audio into text.

P-Video 2 Draft is a faster, lower-cost preview of P-Video 2 for experimenting before the full-quality run.

P-Video 2 is the quality-focused successor to P-Video, with native-speech lipsync and stronger identity consistency.

Seedance 2.0 Mini A smaller version of Seedance that is probably good enough for most purposes Less censorship than the full version

Kling Video 3.0 Omni: Unified multimodal video generation with reference images, video editing, native audio, and multi-shot control

ByteDance's multimodal video generation model with native audio, multimodal reference inputs, and intelligent duration control.

Wan 2.1 14B text to video in 480p is fast and flexible

P-Video is a versatile model that allows you to add input audio and get output audio with lipsync.

P-Video Draft is a high-speed, cost-effective generation model that allows you to experiment until you are ready to use the full P-Video.

Seedance 2.0 Fast is a specialized text-to-video version of the model, designed for speed and audio integration.

Vidu Q1 is a high-definition model optimized for 1080p clarity and realistic motion.

MiniMax Hailou 2.3 Fast provides rapid video generation with highly customizable duration and quality settings.

Pixverse v. 5.6 is a versatile and cost-effective model.