AI Models

EasySite uses multiple AI models to generate your websites. Each model has different strengths.

FreeFree Tier

High-quality open-source models, free forever. Great for most websites.

Auto

Context: —

Automatically picks the best available model for your plan. Handles rate limits by switching models.

Best for: Recommended for most users. No need to think about model selection.

Qwen3 Coder

Context: 262K

Purpose-built coding model by Alibaba. 480B MoE architecture with 35B active parameters.

Best for: Best free model for code generation. Excellent at structured multi-file output.

Qwen 3.6 Plus

Context: 1M

Latest general-purpose model from Alibaba with massive context window.

Best for: Strong at both creative design and code. Great for complex multi-page sites.

Gemma 3 27B

Context: 131K

Google's open-weight model, 27B parameters. Reliable and consistent.

Best for: Consistent structured output. Good at following format instructions.

Llama 3.3 70B

Context: 65K

Meta's 70B parameter model. Strong general-purpose capabilities.

Best for: Well-rounded. Good at responsive design and modern CSS.

Nemotron 3 Super

Context: 262K

NVIDIA's 120B MoE model with 12B active parameters. Large context window.

Best for: Good general-purpose coding. Handles large projects well.

Hermes 3 405B

Context: 131K

Nous Research fine-tune of Llama 3.1 405B. The largest free model available.

Best for: Highest quality among free models. Best for complex multi-page sites.

GLM-4.7 Flash

Context: 128K

ZhipuAI's fast model. Direct API access, not routed through OpenRouter.

Best for: Always available. Good fallback when OpenRouter models are rate-limited.

Qwen 3 235B (Cerebras)

Context: 128K

Qwen 3 235B running on Cerebras hardware. Ultra-fast inference.

Best for: Fastest inference speed. Great for code generation and complex tasks.

Llama 3.1 8B (Cerebras)

Context: 128K

Llama 3.1 8B on Cerebras. Lightning fast for simple tasks.

Best for: Instant responses. Good for quick edits and simple sites.

ProPro Tier

Premium models from Anthropic, Google, and OpenAI. Better design quality and instruction following.

Claude Sonnet 4

Context: 200K

Anthropic's balanced model. Excellent reasoning with fast response times.

Best for: Best balance of speed and quality. Exceptional at clean, semantic HTML and accessibility.

Gemini 2.5 Pro

Context: 1M

Google's most capable model with massive context and multimodal abilities.

Best for: Excels at complex layouts, animations, and multi-page sites.

GPT-4o

Context: 128K

OpenAI's flagship multimodal model. Fast and versatile.

Best for: Great at creative design and modern UI patterns. Strong JavaScript generation.

UnlimitedUnlimited Tier

The most powerful models available. Best for complex, production-ready sites.

Claude Opus 4

Context: 200K

Anthropic's most powerful model. Top-tier reasoning and code generation.

Best for: Highest quality output. Best for production sites with complex interactions.

OpenAI o3

Context: 200K

OpenAI's advanced reasoning model with deep chain-of-thought capabilities.

Best for: Deep reasoning for complex architectures. Excellent problem-solving.

How Model Selection Works

Choose a model from the dropdown in the editor. If a model is unavailable, EasySite automatically falls back to the next best option. Free users get access to 4 powerful open-source models. Upgrade to Pro or Unlimited for premium models.

Tips for Best Results

  • 1Be specific about colors, layout, and content in your prompt.
  • 2For complex sites, use Pro or Unlimited models for better quality.
  • 3Say "continue" if the AI output gets cut off mid-generation.
  • 4Use iterative prompts — start simple, then refine with follow-up requests.
Start Building