About the AI Models Directory
We track 466 models from leading providers including Google, NVIDIA, xAI, OpenAI, Kuaishou, and more. Each model profile includes benchmark scores across general intelligence, coding, math, reasoning, speed, and cost metrics.
Models are categorized as Flagship, Mid-Range, Budget, or Open Source based on their capability tier and pricing. Click any model to view its full benchmark profile, or use the Compare tool to see side-by-side comparisons, or check Pricing for detailed cost analysis.
Frequently Asked Questions
How many AI models do you track?
We currently track 466 large language models from 41 leading providers including OpenAI, Anthropic, Google, Meta, DeepSeek, xAI, Qwen, and Mistral. New models are added as they launch.
What is the difference between Flagship, Mid-Range, Budget, and Open Source models?
Flagship models (e.g. GPT-5.2, Claude Opus 4.6) offer peak capability at premium prices. Mid-Range models balance quality and cost. Budget models (e.g. GPT-5 Nano) prioritize low cost for high-volume use. Open Source models (e.g. Llama, Qwen) can be self-hosted and fine-tuned freely.
Which AI provider has the most models?
OpenAI and Google currently offer the largest model lineups, each with 8+ models spanning flagship to budget tiers. Anthropic, Meta, and DeepSeek each offer 4-6 models, while xAI, Qwen, and Mistral round out the directory.
How often is the AI models directory updated?
The directory is updated within days of a new model launch or pricing change. Benchmark scores are refreshed as new evaluation results become available from official leaderboards and independent testing platforms.
What data is shown on each model profile?
Each model profile includes Chatbot Arena ELO, SWE-bench Verified, MMLU-Pro, HumanEval, and 20+ other benchmark scores, plus input and output pricing per 1M tokens, output speed, context window size, and provider details.