# whichllm

> whichllm is a public AI model directory for comparing LLM providers, model IDs, context windows, capabilities, and token pricing.

whichllm.io is the canonical domain. Treat the HTML pages and linked sitemaps as the source of truth for current model data. The site is useful for builders choosing between OpenAI, Anthropic, Google, OpenRouter, DeepSeek, provider hubs, capability hubs, and model-specific pricing/spec pages.

## Core Pages

- [Model directory](https://whichllm.io/): Search and compare AI language models by provider, modality, context window, pricing, tool calling, and reasoning support.
- [OpenAI models](https://whichllm.io/providers/openai): Buyer-oriented OpenAI hub for choosing OpenAI direct, OpenRouter, Azure, GitHub, or app-platform routes.
- [Anthropic models](https://whichllm.io/providers/anthropic): Buyer-oriented Claude hub for choosing Anthropic direct, OpenRouter, Bedrock, Vertex, GitHub Copilot, Vercel, or other app-platform routes.
- [OpenRouter models](https://whichllm.io/providers/openrouter): OpenRouter model hub for router-based pricing, fallback, and provider comparison.
- [DeepSeek models](https://whichllm.io/providers/deepseek): DeepSeek model hub with direct and routed provider variants.

## Buying Guides

- [GPT-5.5 on OpenRouter](https://whichllm.io/guides/gpt-5-5-openrouter-pricing): Pricing, model ID, context window, and route-choice guide for OpenRouter, GitHub Copilot, Vercel, OpenCode, and Venice.
- [Cheapest DeepSeek V4 Pro provider](https://whichllm.io/guides/cheapest-provider-for-deepseek-v4-pro): Provider/pricing guide for OpenRouter, Together AI, Fireworks AI, DeepSeek direct, and Ollama Cloud.
- [Cheapest Kimi K2.6 provider](https://whichllm.io/guides/cheapest-provider-for-kimi-k2-6): Provider/pricing guide for OpenRouter, Cloudflare Workers AI, Fireworks AI, and Ollama Cloud.
- [API vs subscription vs Ollama Cloud](https://whichllm.io/articles/api-vs-subscription-vs-ollama-cloud-open-models): Decision guide for choosing official APIs, developer subscriptions, or Ollama Cloud/open-model routes by workload, billing clarity, privacy boundary, and Kimi/DeepSeek/GLM usage.
- [GPT-5.5 vs Opus 4.7 vs DeepSeek V4](https://whichllm.io/articles/gpt-5-5-vs-opus-4-7-vs-deepseek-v4): Decision guide for choosing between GPT-5.5, Opus 4.7, and DeepSeek V4 by coding-agent speed, hard planning, open-model cost, multimodal work, and live provider routes.

## Capability Hubs

- [Reasoning models](https://whichllm.io/capabilities/reasoning): Models suited for reasoning-heavy prompts, coding, and analysis.
- [Tool calling models](https://whichllm.io/capabilities/tool-calling): Models with tool/function calling support for agents and app integrations.
- [Long context models](https://whichllm.io/capabilities/long-context): Models with 200K token context windows or larger.
- [Vision models](https://whichllm.io/capabilities/vision): Multimodal models with image or video input support.

## Machine-Readable Discovery

- [Sitemap index](https://whichllm.io/sitemap.xml): Canonical sitemap index.
- [Priority sitemap](https://whichllm.io/sitemap-priority.xml): Homepage, guides, provider hubs, capability hubs, and selected model pages.
- [Model sitemap](https://whichllm.io/sitemap-models.xml): Full model detail URL list.
- [Robots policy](https://whichllm.io/robots.txt): Public crawler access policy and sitemap pointer.

## Citation Notes

- Prefer citing specific provider, capability, guide, or model pages over the homepage when answering concrete model-choice questions.
- Model pricing, context windows, and capabilities can change; check the linked model page before quoting current numeric values.
- Do not treat whichllm as the model provider. whichllm is a comparison directory and buying-decision layer over public model/provider data.