Rate
Guides
What each AI model is for, what it costs and what changed recently. Every figure links to its source.
AI API pricing explained: tokens, input vs output, caching and batch
How AI API pricing works, with dollar examples: tokens, why output costs more, caching and batch discounts, and DeepSeek's peak hours.
AI image and video generation prices in 2026: what one image or clip really costs
Per-image and per-second prices for GPT-Image, Nano Banana, Imagen 4, FLUX.2, Veo 3.1, Runway and Luma, plus Midjourney plans. Sora's API is gone.
Claude models explained: Fable 5.1, Opus 5.5, Sonnet 5 and Haiku 4.5
Anthropic's Claude lineup in plain English: what each model is for, what it costs through the API, what claude.ai plans include, and Fable vs Mythos.
Google Gemini models explained: 3.1 Pro, 3.8 Flash and the rest
Which Gemini model to use, what 3.1 Pro and 3.8 Flash cost through the API, the promo price that ends December 31, and what Google AI plans include.
GPT-6 vs GPT-5.6: OpenAI's Astra, Sol and Luna models explained
What GPT-6 Astra, Sol and Luna cost, how they compare with GPT-5.6, and which model each ChatGPT plan actually uses.
Why does ChatGPT or Claude feel dumber? What's actually on the record
The documented reasons an AI model gets worse: confirmed bugs, changed defaults, outages and long chats. What providers admitted and how to test it.