Model Library

Choose from ready-to-go models -- Or bring your own custom model.

Official models

Official Model Library

Explore deployable models curated for Vast.ai, with recommended engines and hardware configurations.

What this view showsModel details, deployment engines, and hardware recommendations come directly from the official catalog.

ModelEngineRecommended GPUAction

Qwen Image 2.1Alibaba

EngineComfyui

Recommended GPU

RTX 4090Supply unavailable

View model

Qwen3.8 27BAlibaba

EngineVllm

Recommended GPU

RTX PRO 6000 SSupply unavailable

View model

MiniMax H3MiniMaxAI

EngineComfyui

Recommended GPU

RTX PRO 6000 SSupply unavailable

View model

Kimi K3Moonshot AI

EngineVllm

Recommended GPU

8× B300Supply unavailable

View model

Krea 2 TurboKrea

EngineComfyui

Recommended GPU

RTX 4090Supply unavailable

View model

GLM 5.2Z.ai

EngineVllm

Recommended GPU

8× H200Supply unavailable

View model

Qwen3.5 35B A3BAlibaba

EngineVllm

Recommended GPU

RTX PRO 6000 SSupply unavailable

View model

Qwen3.6 35B A3BAlibaba

EngineVllm

Recommended GPU

RTX PRO 6000 SSupply unavailable

View model

Source: Official Vast.ai Model Library.Official catalog

Featured Models

image

Qwen Image 2.1

Compact 7B unified text-to-image and image editing model in the Qwen series with native transparent (RGBA) generation and multi-reference editing

textvision

Qwen3.8 27B

Dense 27B vision-language model with hybrid Gated DeltaNet attention and 262K context

video

MiniMax H3

Audio-video foundation model generating up to 2K 24fps clips with natively synchronized stereo audio from text, image, or reference inputs

textvision

Kimi K3

Moonshot AI's 2.8T-parameter multimodal MoE — full-precision vLLM flagship on 8xB300, plus single-node Unsloth 1-bit and 2-bit dynamic GGUF quants

image

Krea 2 Turbo

Distilled few-step version of Krea 2, a 12B text-to-image diffusion transformer for fast high-quality image generation

text

GLM 5.2

753B MoE model with 1M-token context for agentic reasoning, coding, and tool use

Deploy Your Own

Looking for PyTorch? Vast.ai has thousands of templates to get started immediately, or you can create your own template specifically for your needs. Explore or start building your own template.

Deploy Your Own

Image Generation

Text to image

Qwen Image 2.1

Compact 7B unified text-to-image and image editing model in the Qwen series with native transparent (RGBA) generation and multi-reference editing

Krea 2 Turbo

Distilled few-step version of Krea 2, a 12B text-to-image diffusion transformer for fast high-quality image generation

Seedream 5.0 Pro

ByteDance Seedream 5.0 Pro image generation, served as a Vast Model Endpoint

HiDream I1 Full

Open-source image generative foundation model with 17B parameters that achieves state-of-the-art image generation quality within seconds

Video Generation

Make and edit video from text or images

MiniMax H3

Audio-video foundation model generating up to 2K 24fps clips with natively synchronized stereo audio from text, image, or reference inputs

EmbeddingGemma 2

Google's multimodal embedding model mapping text, code, images, video, and audio into one 768-dimensional vector space, served as an OpenAI-compatible embeddings API

Mochi 1 Preview

Mochi 1 preview is an open state-of-the-art video generation model with high-fidelity motion and strong prompt adherence in preliminary evaluation.

Wan2.2 I2V A14B (FP8)

Wan2.2 introduces a Mixture-of-Experts (MoE) architecture into image to video diffusion models

Audio Generation

Make speech, music, and sound effects

EmbeddingGemma 2

Google's multimodal embedding model mapping text, code, images, video, and audio into one 768-dimensional vector space, served as an OpenAI-compatible embeddings API

VoxCPM2

OpenBMB's tokenizer-free 48 kHz text-to-speech model with 30 languages, voice design from a text description, and voice cloning

Voxtral Mini 4B Realtime

Multilingual streaming speech-to-text model from Mistral AI that transcribes live audio in 13 languages with under 500 ms of delay

Chatterbox

Resemble AI's production-grade open TTS with zero-shot voice cloning, emotion exaggeration control, and multilingual synthesis across 23 languages

Text Generation

Language models for chat, code, and analysis

Qwen3.8 27B

Dense 27B vision-language model with hybrid Gated DeltaNet attention and 262K context

Kimi K3

Moonshot AI's 2.8T-parameter multimodal MoE — full-precision vLLM flagship on 8xB300, plus single-node Unsloth 1-bit and 2-bit dynamic GGUF quants

GLM 5.2

753B MoE model with 1M-token context for agentic reasoning, coding, and tool use

Qwen3.5 35B A3B

Efficient 35B MoE with 3B active params, unified vision-language reasoning

Multimodal

Understand text, images, and audio together

MiMo V2.6 Pro RL

Xiaomi's flagship MoE reasoning model - 1.02T total / 42B active params, trained with large-scale mixed RL for coding, agents and cybersecurity, with a 1M-token context window

Inkling-Small

Thinking Machines' 276B MoE (12B active) that takes text, images and audio, with a 1M-token context for coding, agents and reasoning

MiMo V2.6 Flash RL

Xiaomi's 309B-parameter Mixture-of-Experts model (15B active) for agentic coding, tool use, and cybersecurity, with a 1M-token context window

Gemma 4 12B IT

Gemma 4 12B Unified encoder-free omni model by Google with native text, image, and audio input and 256K context

Computer Vision

Describe, analyze, and read text from images

Qwen3.8 27B

Dense 27B vision-language model with hybrid Gated DeltaNet attention and 262K context

Kimi K3

Moonshot AI's 2.8T-parameter multimodal MoE — full-precision vLLM flagship on 8xB300, plus single-node Unsloth 1-bit and 2-bit dynamic GGUF quants

Qwen3.5 35B A3B

Efficient 35B MoE with 3B active params, unified vision-language reasoning

Qwen3.6 35B A3B

Agentic coding MoE with hybrid Gated DeltaNet and vision support

Model Library

Official Model Library

Featured Models

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen Image 2.1

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.8 27B

MiniMax H3

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Kimi K3

Krea 2 Turbo

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}GLM 5.2

Deploy Your Own

Image Generation

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen Image 2.1

Krea 2 Turbo

Seedream 5.0 Pro

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}HiDream I1 Full

Video Generation

MiniMax H3

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}EmbeddingGemma 2

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Mochi 1 Preview

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Wan2.2 I2V A14B (FP8)

Audio Generation

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}EmbeddingGemma 2

VoxCPM2

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Voxtral Mini 4B Realtime

Chatterbox

Text Generation

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.8 27B

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Kimi K3

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}GLM 5.2

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.5 35B A3B

Multimodal

MiMo V2.6 Pro RL

Inkling-Small

MiMo V2.6 Flash RL

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Gemma 4 12B IT

Computer Vision

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.8 27B

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Kimi K3

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.5 35B A3B

.css-lw4uyf{min-height:32px;max-height:32px;min-width:32px;max-width:32px;aspect-ratio:1/1;box-sizing:content-box;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;margin:0;color:#e9e9e9;}Qwen3.6 35B A3B

Qwen Image 2.1

Qwen3.8 27B

Kimi K3

GLM 5.2

Qwen Image 2.1

HiDream I1 Full

EmbeddingGemma 2

Mochi 1 Preview

Wan2.2 I2V A14B (FP8)

EmbeddingGemma 2

Voxtral Mini 4B Realtime

Qwen3.8 27B

Kimi K3

GLM 5.2

Qwen3.5 35B A3B

Gemma 4 12B IT

Qwen3.8 27B

Kimi K3

Qwen3.5 35B A3B

Qwen3.6 35B A3B