Auto
Built by Magai · 1 version
OpenRouter Auto Router selects the best available chat model for each prompt.
The full AI models list, checked against the app daily
Every AI model in Magai, and every version
ChatGPT, Claude, Gemini and Grok for chat, GPT Image, Nano Banana, Veo and Kling Video for pictures and film, and the rest of the list below, all in one chat. What you see here is what you get when you sign in.
- models
- 81
- providers
- 24
- for chat
- 57
- for image and video
- 24
Last checked
Select AI Model
Choose from available AI models with different context limits and usage multipliers.
All Capabilities
Chat models
Auto
ChatGPT
Claude
Gemini
Grok
DeepSeek
GLM
Kimi
Llama
MiMo
MiniMax
Mistral
Muse
Nemotron
Nova
Perplexity
Voice GPT
Image models
GPT Image
Nano Banana
Flux Image
Grok Imagine Image
Ideogram
Kling Image
Leonardo.ai
Luma UNI-1
Muse Image
Qwen Image
Recraft
Reve
Runway Image
Seedream
Video models
Veo
Kling Video
Runway Video
Flux Video
Gemini Omni Flash
Grok Imagine Video
Leonardo Motion
Luma
MiniMax
Seedance
The picker in the app, with today’s list. Try the search.
Chat models 57
For writing, research, code and anything you would type to an assistant. Switch between them mid-conversation with a slash; each one reads the whole thread.
ChatGPT
By OpenAI · 16 versions
OpenAI GPT-6 Luna is a fast, cost-efficient GPT-6 model for high-volume and latency-sensitive workloads. 1.05M token context; accepts text and image inputs.
OpenAI GPT-6 Luna Pro is GPT-6 Luna served with pro reasoning for higher-quality responses on complex tasks. 1.05M token context; accepts text and image inputs.
OpenAI GPT-6 Sol is a cost-efficient high-end GPT-6 model for professional work, agentic coding, and long-horizon software engineering. 1.05M token context; accepts text and image inputs.
OpenAI GPT-6 Sol Pro is GPT-6 Sol served with pro reasoning for higher-quality responses on complex tasks. 1.05M token context; accepts text and image inputs.
OpenAI gpt-oss-120b is an open-weight mixture-of-experts model for high-reasoning, agentic, and general-purpose use. 131K token context.
OpenAI GPT-5.4 Mini is a faster GPT-5.4 variant for high-throughput reasoning, coding, and chat. 400K token context; accepts text and image inputs.
OpenAI GPT-5.4 Nano is the most lightweight GPT-5.4 variant for speed-critical and high-volume tasks. 400K token context; accepts text and image inputs.
GPT-5.5 – OpenAI's frontier model for complex professional workloads. Builds on GPT-5.4 with stronger reasoning, higher reliability, and improved token efficiency. 1M+ context window, text and image inputs.
GPT-5.5 Pro - OpenAI's high-capability model optimized for deep reasoning, accuracy, and complex professional workloads. 1M+ context window, text and image inputs.
OpenAI GPT-5.6 Luna is a fast, cost-efficient model for high-volume chat, classification, and lightweight agentic workflows. 1.05M token context; accepts text and image inputs.
OpenAI GPT-5.6 Luna Pro is GPT-5.6 Luna served with pro reasoning for higher-quality responses on complex tasks. 1.05M token context; accepts text and image inputs.
GPT-5.6 Sol – the flagship GPT-5.6 model for complex reasoning, coding, and long-horizon agentic workflows. 1.05M token context window; accepts text, image, and file inputs.
GPT-5.6 Sol Pro – GPT-5.6 Sol served with reasoning mode set to pro for higher-quality responses on complex tasks. 1.05M token context window; accepts text, image, and file inputs.
OpenAI GPT-5.6 Terra is a balanced model for everyday coding, reasoning, and agentic work. 1.05M token context; accepts text and image inputs.
GPT-5.6 Terra Pro – GPT-5.6 Terra served with reasoning mode set to pro for higher-quality responses on complex tasks. 1.05M token context window; accepts text, image, and file inputs.
GPT-6 Astra – OpenAI's flagship model for demanding end-to-end work, suited for advanced analysis, software engineering, deep research, and long-horizon agentic tasks. 1.05M token context window; accepts text, image, and file inputs.
Claude
By Anthropic · 6 versions
Claude Sonnet 5.5 – Anthropic's Sonnet-class model for well-scoped everyday work, strong at building features, fixing bugs, and producing polished documents. 1000K token context window; accepts text and image inputs.
Anthropic Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work. 1M token context; accepts text and image inputs.
Anthropic Claude Haiku 4.5 is Anthropic's fastest Claude model, delivering near-frontier intelligence at low latency. 200K token context.
Claude Opus 4.8 – Anthropic's most capable generally available model in the Opus family. Supports text, image, and file inputs with reasoning support and a 1M-token context window. Suited for highly autonomous agents, long-horizon agentic work, knowledge work, and memory-driven tasks. Particularly strong on multi-step reasoning, complex coding, and end-to-end project orchestration across large codebases and long-running async pipelines.
Claude Opus 5 – Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work. 1M token context window; accepts text and image inputs.
Anthropic Claude Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. 1M token context.
Gemini
By Google · 4 versions
Google Gemini 3.1 Pro is Google's frontier reasoning model for software engineering, agentic reliability, and complex multimodal workflows. 1M token context.
Google Gemini 3.5 Flash-Lite is a high-efficiency model with upgraded agentic capabilities, suited for focused subagent work. 1M token context; accepts text and image inputs.
Google Gemini 3.7 Flash is Google's multimodal workhorse for fast agentic workflows, coding, and complex multi-step reasoning. 1M token context; accepts text and image inputs.
Google Gemini 3.8 Flash is Google's most intelligent Flash model, with gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning. 1M token context; accepts text and image inputs.
Grok
By xAI · 3 versions
xAI Grok 4.7 is xAI's flagship model for coding, agentic tasks, and knowledge work. 500K token context; accepts text and image inputs.
xAI Grok 4.5 delivers frontier performance on coding, knowledge work, and STEM. 500K token context; accepts text and image inputs.
xAI Grok 4.6 delivers frontier performance on coding, knowledge work, and STEM. 500K token context; accepts text and image inputs.
DeepSeek
By DeepSeek · 3 versions
DeepSeek V4 Flash is a sparse mixture-of-experts model suited for coding, reasoning, and agent workflows. 1M token context.
DeepSeek V4 Pro is a large-scale mixture-of-experts model for advanced reasoning, coding, and long-horizon agent workflows. 1M token context.
DeepSeek V4.1 Flash is a sparse mixture-of-experts model and the first built on DeepSeek's Causal Encoder-Decoder architecture. 1M token context; accepts text and image inputs.
GLM
By Z.AI · 4 versions
Z.AI GLM-5 is Z.AI's flagship foundation model for complex systems design and long-horizon agent workflows. 202K token context.
Z.AI GLM-5 Turbo is optimized for fast inference in agent-driven workflows. 202K token context.
Z.AI GLM 5.2 is a large-scale reasoning model for long-horizon agent workflows and project-level software engineering. 1M token context.
Z.AI GLM-5.3 is a large-scale reasoning model for complex software engineering and long-horizon agent tasks. 1M token context.
Kimi
By Moonshot AI · 2 versions
Moonshot AI Kimi K2.5 is a native multimodal model with strong visual coding and agent workflows. 262K token context.
Kimi K3 – 2.8T parameter open-weight multimodal reasoning model from Moonshot AI, suited for complex coding, knowledge work, and long-horizon agentic workflows. 1M token context window; accepts text and image inputs.
Llama
By Meta · 2 versions
Meta Llama 4 Maverick is a high-capacity multimodal mixture-of-experts model. 1M token context.
Meta Llama 4 Scout is a mixture-of-experts language model with native multimodal input. 328K token context.
MiMo
By Xiaomi · 2 versions
262k Context, Unmoderated
1.05M Context, Unmoderated
MiniMax
By MiniMax · 2 versions
MiniMax M2.7 is a large language model for autonomous productivity and multi-agent workflows. 205K token context.
MiniMax M3 is a multimodal foundation model for long-horizon agentic work and coding. 1M token context.
Mistral
By Mistral · 3 versions
Mistral Large 3 is an open-weight multimodal mixture-of-experts model for chat, agents, and long-context work. 262K token context.
Mistral Pixtral Large is a multimodal model for documents, charts, and natural images while keeping strong text performance. 128K token context.
Mistral Small 4 unifies flagship Mistral capabilities into a single system with strong reasoning. 262K token context.
Muse
By Meta · 1 version
Meta Muse Spark 1.3 is a multimodal reasoning model for long-running agentic, multi-agent, and coding workflows. 1M token context; accepts text and image inputs.
Nemotron
By NVIDIA · 1 version
NVIDIA Nemotron 3 Nano is a small mixture-of-experts model for efficient specialized agentic systems. 262K token context.
Nova
By Amazon · 2 versions
Amazon Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads. 1M token context.
Amazon Nova Pro is a capable multimodal model balancing accuracy, speed, and cost. 300K token context.
Perplexity
By Perplexity · 4 versions
Perplexity Sonar Deep Research is a research-focused model for multi-step retrieval, synthesis, and reasoning. 200K token context.
Perplexity Sonar is a lightweight search-backed chat model with citations. 127K token context.
200k Context, Online
Perplexity Sonar Pro Search is Perplexity's advanced agentic search system for deeper reasoning and analysis. 200K token context.
Voice GPT
Built by Magai · 1 version
Image models 14
For pictures, from a prompt or from pictures you already have, made in the same thread you were writing in.
GPT Image
By OpenAI · 3 versions
OpenAI GPT Image generates and edits production-ready images with strong instruction-following, layout control, and text rendering, including reference-based edits.
Edit ModeImage to ImageMulti Images to ImageText to Image
Versions
- v2
- 2.5 Flare
- 2.5 Sunburst
Nano Banana
By Google · 4 versions
Google Gemini Image (Nano Banana) generates and edits images conversationally from text and image inputs, with native multimodal understanding.
Edit ModeImage to ImageMulti Images to ImageText to Image
Versions
- v2
- 2 Lite
- v1 Standard
- Lite
Flux Image
By Black Forest Labs · 4 versions
Black Forest Labs FLUX.2 generates and edits photorealistic images with multi-reference control, precise color, and production-ready text, at up to 4 megapixels.
Edit ModeImage to ImageMulti Images to ImageText to Image
Versions
- Kontext Max
- 2 Standard
- 2 Max
- 2 Klein
Grok Imagine Image
By xAI · 2 versions
xAI Grok Imagine generates and edits images from text and references, with photorealistic scenes, logos, and precise prompt following.
Edit ModeImage to ImageText to Image
Versions
- Image
- Image 2.0
Ideogram
By Ideogram · 2 versions
Ideogram generates photorealistic images with accurate prompt alignment, legible text, and production-ready layouts for design and brand work.
Text to Image
Versions
- v4
- v3
Kling Image
By Kling
Kling IMAGE generates and edits images from text and references, with native 2K and 4K output and support for multiple reference images.
Version
- v3 Image
Leonardo.ai
By Leonardo
Leonardo Phoenix generates images with strong prompt adherence, visual fidelity, and coherent in-image text.
Image to ImageText to Image
Version
- Standard
Luma UNI-1
By Luma
Luma UNI-1 is a multimodal reasoning image model for text-to-image, precise natural-language edits, and reference-guided generation.
Image to ImageText to Image
Version
- UNI-1
Muse Image
By Meta
Meta Muse Image generates and edits images with faithful instruction-following, accurate text, and support for multiple reference images.
Image to ImageMulti Images to ImageText to Image
Version
- Image
Qwen Image
By Alibaba
Alibaba Qwen Image generates and edits images with strong text rendering and prompt adherence, including reference-based edits.
Image to ImageMulti Images to ImageText to Image
Version
- v3
Recraft
By Recraft · 3 versions
Recraft generates production-ready raster images and native editable vector graphics from text, including logos, icons, and brand assets.
Image to ImageText to Image
Versions
- v3
- v4
- v4.1 Vector
Reve
By Reve · 2 versions
Reve generates high-quality images with strong prompt adherence, layout intelligence, and accurate text rendering. Covers text-to-image, image editing, and multi-image remix.
Image to ImageMulti Images to ImageText to Image
Versions
- 2.1
- Standard
Runway Image
By Runway
Runway Gen-4 Image creates new images from a reference image, preserving style and composition while following a prompt.
Image to Image
Version
- v4 Image
Seedream
By ByteDance · 2 versions
ByteDance Seedream generates and edits images in a unified model, with strong prompt following, text rendering, and high-resolution output.
Edit ModeImage to ImageText to Image
Versions
- v5
- v4
Video models 10
For clips, from a prompt, from an image, or from footage you upload.
Veo
By Google · 2 versions
Google DeepMind Veo generates cinematic video with native synchronized audio from text and images, including reference guidance and video extension.
Image to VideoText to Video
Versions
- v3
- v3.1
Kling Video
By Kling · 3 versions
Kling VIDEO generates video from text and images with native audio, multi-shot narrative control, and clips up to 15 seconds.
Image to VideoText to VideoVideo to Video
Versions
- O3
- V3
- V3 Turbo
Runway Video
By Runway · 3 versions
Runway Gen-4.5 generates cinematic video from text or images, with precise prompt adherence, realistic motion, and camera control.
Image to VideoVideo to Video
Versions
- v4
- Aleph 2.0
- v4.5
Flux Video
By Black Forest Labs
Black Forest Labs FLUX 3 generates video with synchronized native audio from text, images, or keyframes, including clips up to 20 seconds.
Edit ModeVideo to Video
Version
- 3
Gemini Omni Flash
By Google · 2 versions
Google Gemini Omni Flash is a fast video generation and conversational editing model. It covers text-to-video, image-to-video, video editing, and reference-to-video, with synchronized audio.
Image to VideoText to VideoVideo to Video
Versions
- Omni Flash
- Omni Flash 1.1
Grok Imagine Video
By xAI
xAI Grok Imagine generates cinematic video with native audio from text, images, or references, including edit, extend, and clips up to 15 seconds.
Edit ModeImage to VideoText to VideoVideo to Video
Version
- Video
Leonardo Motion
By Leonardo · 2 versions
Leonardo Motion generates short videos from text prompts or still images, with motion control and frame interpolation.
Image to VideoText to Video
Versions
- v1
- v2
Luma
By Luma · 2 versions
Luma Ray generates realistic video with coherent motion from text or images.
Image to VideoText to Video
Versions
- Ray 2
- Ray 3.2
MiniMax
By MiniMax · 3 versions
MiniMax Hailuo generates cinematic video from text or images, with strong instruction following and physics-aware motion.
Image to VideoText to VideoVideo to Video
Versions
- Hailuo 02
- H3
- H3 Max
Seedance
By ByteDance · 3 versions
ByteDance Seedance generates video from text and images, with multi-shot storytelling, stable motion, and multimodal audio-video generation.
Image to VideoText to VideoVideo to Video
Versions
- v2.5
- v2
- Lite
One conversation
The list is not a menu of apps. It is one chat.
Type a slash and a model's name, and the next reply comes from it. Every model reads the whole thread, so Claude's draft, Gemini's fact check and the picture Nano Banana makes next all build on the same work. Nothing is copied between tabs.
You
ClaudeYou
Gemini/Nano Banana A header image for it: linen on a sunlit table, soft greens.
GeminiQuestions about the models
81 today: 57 chat models, 14 image models and 10 video models, from 24 providers. Counting every version of every image and video model, that is 107 choices in the picker. The number moves as models ship and retire, which is why this page reads it from the app every day.
16: GPT-6 Luna, GPT-6 Luna Pro, GPT-6 Sol, GPT-6 Sol Pro, GPT OSS 120B, GPT-5.4 Mini, GPT-5.4 Nano, GPT-5.5, GPT-5.5 Pro, GPT-5.6 Luna, GPT-5.6 Luna Pro, GPT-5.6 Sol, GPT-5.6 Sol Pro, GPT-5.6 Terra, GPT-5.6 Terra Pro and GPT-6 Astra. They sit together under ChatGPT in the model picker, and you pick the version from its menu. See each one.
6: Claude Sonnet 5.5, Claude Opus 5.5, Claude Haiku 4.5, Claude Opus 4.8, Claude Opus 5 and Claude Sonnet 5. They sit together under Claude in the model picker, and you pick the version from its menu. See each one.
4: Gemini 3.1 Pro, Gemini 3.5 Flash-Lite, Gemini 3.7 Flash and Gemini 3.8 Flash. They sit together under Gemini in the model picker, and you pick the version from its menu. See each one.
3: Grok 4.7, Grok 4.5 and Grok 4.6. They sit together under Grok in the model picker, and you pick the version from its menu. See each one.
GPT Image, Nano Banana, Flux Image, Grok Imagine Image, Ideogram, Kling Image, Leonardo.ai, Luma UNI-1, Muse Image, Qwen Image, Recraft, Reve, Runway Image and Seedream. Each one works in the same thread as the chat, so you can write the brief with one model and make the picture with another.
Veo, Kling Video, Runway Video, Flux Video, Gemini Omni Flash, Grok Imagine Video, Leonardo Motion, Luma, MiniMax and Seedance. Most take a prompt or a still image; several also take a clip you upload.
Yes. Magai talks to each lab's own API on your behalf, so you get the current models from OpenAI, Anthropic, Google, xAI, DeepSeek, Meta and the rest, not a copy or a cheaper stand-in.
No. One Magai subscription is the only account you need; Magai calls every provider for you.
Standard includes every standard model. Pro and Ultra add the premium models, the most expensive tiers such as the top GPT and Claude versions. Compare the plans.
Yes. Type a slash and the model's name, and the next reply comes from that model, which reads the whole conversation so far. Nothing is pasted or retyped.
As fast as we can add them. When a major provider ships an upgrade we implement it, and this list picks it up from the app within a day. It was last checked on September 29, 2026.
Every model on this page, in one chat.
Start with the one you know. Switch to any other mid-sentence.
From $20 a month. 30-day money-back guarantee.