Looking for the best AI model? Our AI model tier list ranks the top language models and image generators by community preference. Updated weekly from Arena AI blind user votes, each model earns an S, A, B, C, or D tier based on ELO scores. S tier represents the top 5% — the absolute best AI models available today. Whether you need the best AI for coding, the top image generation model, or a general-purpose LLM, our AI model ranking helps you compare and choose.
Community-ranked AI models by category. Based on Arena AI ELO scores, updated weekly.
Top: Anthropic: Claude Fable 5
Top: Anthropic: Claude Fable 5
Top: Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)
Top: Anthropic: Claude Fable 5
Top: Anthropic: Claude Fable 5
The top-ranked model in our general AI tier list is Anthropic: Claude Fable 5 with an ELO score from Arena AI blind user votes. Our S/A/B/C/D tier system ranks models from best to worst based on community preference. S tier represents the top 5% of all AI models.
Models are ranked using ELO scores from Arena AI (formerly LMSYS Chatbot Arena), where real users vote in blind comparisons. S tier = top 5%, A tier = top 20%, B tier = top 50%, C tier = top 80%, D tier = bottom 20%. Data is fetched weekly from Arena AI and OpenRouter APIs and the rankings recalculated automatically.
The coding tier list ranks AI coding assistants by ELO score. Each model earns an S/A/B/C/D tier based on its performance in blind code generation comparisons. Visit the coding tier page to see the full ranking of best AI models for programming and software development.
The image generation tier list ranks AI image models including GPT Image, Stable Diffusion, Flux, Midjourney, and others by community preference ELO scores. Check the image generation tier page for the latest rankings of the best AI image generators.
New rankings are fetched every Monday from the Arena AI leaderboard and OpenRouter APIs. The site auto-deploys with the latest data, so tier lists are never more than a week old. Rankings shift as new models are released and existing models improve.
S tier = top 5% of models. A tier = next 15% (top 5-20%). B tier = next 30% (top 20-50%). C tier = next 30% (top 50-80%). D tier = bottom 20%. Tiers are recalculated each week as new Arena AI data arrives and model rankings shift.