Reference-guided video, audio-driven, up to 30s

Long-horizon agentic reasoning, 1M context

High-fidelity images with near-perfect text rendering

Faster WAN 3.0 Omni at a higher price

Detail-preserving image upscaling up to 4x

Video super-resolution to 4K, precise or creative

SDR-to-HDR video, including AI-generated footage

Detail-preserving video upscaling with softness control

GLM 5.3

Long-horizon coding agents, 1M context

Text, image, and audio to video, up to 4K

Reference or first/last frame video with audio

Text, keyframe and video-continuation generation

Animate images into video with audio

Text-to-video with audio, up to 30s

Efficient MoE for agents, 1M context

MiniMax H3

Open weights omni-modal video with audio

Reference-video control fine-tuning

Agentic coding and vision, 1M context

Efficient high-volume model, 1M context

Frontier reasoning and coding, 1M context

Balanced reasoning and cost, 1M context

Agentic coding with 1M context

Precise image generation and editing

Low-cost 1K image generation

Text-to-speech with voice cloning and prosody control

Krea 2 Raw

Open-weights text-to-image base for LoRA fine-tuning

Text-to-video up to 1080p

GLM 5.2

Coding and agents, 1M-token context

Image-to-video up to 1080p

Reference-to-video, up to 9 images

Adaptive thinking coding and agents

Luma Ray 3.2 - Image to Video

Image-to-video with prompt control, up to 1080p

Luma Ray 3.2 - Reframe

Aspect-ratio reframe with outpainting, up to 1080p

Luma Ray 3.2 - Text to Video

Text-to-video with prompt control, up to 1080p

Luma Ray 3.2 - Video to Video

Prompt-guided video editing, up to 1080p

Lift SDR video into HDR

Image-to-video with audio (Grok 1.5)

Agentic coding, 1M-token context

HDR video conversion

Omni-modal short video generation and editing

Image-to-video up to 1080p

Reference-to-video, up to 9 images

Text-to-video up to 1080p

Prompt-based video editing up to 1080p

Image-to-video, 1080p, frame control

Multi-reference video, up to 1080p

Text-to-video, 1080p, multi-shot

Large MoE for reasoning and coding

Reasoning and coding model, 1M context

Deep reasoning, 1M-token context

Agentic coding, 1M-token context

Prompt-based video editing

Fast, affordable image-to-video

Fast reference-guided video generation

Fast, affordable text-to-video

Animate images into 4k video

Reference-guided video with audio

Cinematic text-to-video, up to 4k

Open multimodal model, 256K context

Lightweight on-device multimodal model

Compact on-device multimodal model

Detail-preserving video upscaling

Fast model for coding, computer use

Consistent depth maps from video

Controllable text/image-to-video generation

Agentic coding model, 1M context

Text, image, and audio to video, up to 4K

Continue a clip from its start or end

Image-to-video with native audio

Convert a clip to a new aspect ratio

Regenerate a section of a clip

Pro quality at Flash speed

Multimodal model with vision, 256K context

Multimodal model with vision, 256K context

Multimodal model with vision, 256K context

Multimodal model with vision, 256K context

Intelligent visual reasoning image model

Multimodal reasoning over text, audio, video

Balanced coding and agents, 1M context

Text-and-image to image generation

Coding and agents, 1M context

Unified O3 video generation and edit

Transfer motion from video to image

Versatile image styles by xAI

Edit images with Grok Imagine

Edit videos with Grok Imagine

Image-to-video with audio by xAI

Ideogram V4

Typography-focused image and design generation

Krea 2 Large

High-fidelity text-to-image with style references

Compact, fast text-to-image model

Compact, fast text-to-image model

Photorealistic portraits and natural scenes

Improved natural-language image editing

Low-latency multimodal model, 1M context

Cinematic image-to-video with audio

Cinematic text-to-video with native audio

True-color precision rendering

Image-to-video with audio and lip-sync

Reference video with character consistency

Reasoning model, 400K context

Cinematic image-to-video generation

Reference-driven cinematic video

Advanced video editing

Text-to-image and editing, up to 4K

Fast photorealistic bilingual image generation

Open text-to-image with multi-reference

Tunable open text-to-image model

Professional FLUX.2 image generation

Software engineering and agentic workflows

Studio-quality 4K image generation

Zero-shot image segmentation

Zero-shot video object segmentation

Adaptive reasoning and instruction following

Professional AI image upscaling

Professional AI video upscaling

Text/image-to-video, audio, narrative control

Fast text/image-to-video with audio, 4K

Cost-efficient text/image-to-video by Google

Realistic video generation with audio

Fast cinematic image-to-video

Multi-image natural-language editing

On-device photorealistic image generation

Balanced coding, agents, computer use

Photorealistic 4K image generation

OpenAI audio-input chat model

Natural-language image editing

Vision-language model, 256K context

Vision-language model, long context

Vision-language model for text and images

Open text-to-image with editing tools

Multi-model routing, 400K context

Low-cost GPT-5 for real-time apps

Ultra-low-latency GPT-5 model

Open MoE reasoning model

Open MoE, runs on consumer hardware

Text-to-image with strong text rendering

High-fidelity open text-to-video, 720p

Efficient 720p video on consumer GPUs

Fast multimodal reasoning, 1M context

In-context image editing model

Text/image-to-video with native audio

Multimodal reasoning and coding, 1M context

Open model, 256K context

Model for edge and offline use

Reasoning for coding, math, science

Fast, low-cost multimodal reasoning

Coding and chat, 1M-token context

Lower-cost GPT-4.1, 1M context

Low-latency GPT-4.1 for classification

Multimodal model, 128K context

Multi-step web research and synthesis

Multi-step reasoning over web search

Multilingual on-device chat model

Text-to-video with bilingual text

Fast web-grounded search, built on Llama

Advanced web-grounded search, 200K context

Efficient model for edge deployment

Vision-language model, 128K context

Fast, low-cost multimodal model

Fine-detail video upscaling

Text embeddings for search, RAG

Multimodal model for text, audio, vision

MoE model for code and multilingual

Long-context reasoning model, 128K

High-quality text embeddings

Efficient text embeddings

Versatile video upscaling

Efficient open MoE model

Face-focused video upscaling

General-purpose open model

Cost-effective image-to-video, up to 720p

Cost-effective reference-guided video generation

Cost-effective text-to-video, up to 720p

Model for on-device and edge

Reasoning for science, math, code

Fast reasoning for STEM tasks

Efficient multilingual model

Coding and agentic work, 1M context

Creative video upscaling for AI video

Generative image upscaling with facial detail

Creative generative image upscaling with prompts

SDR-to-HDR video, 10-bit HDR10 output

Detail-preserving image enhancement and upscaling

All-in-one image enhancement and upscaling

Lightweight text-to-video on consumer GPUs