Blog
Notes from the catalog: price math you can check, benchmark rankings honestly labeled, and the caveats other pricing pages leave out.
The workhorse tier: best cheap models in mid-2026
Real prices and context windows for sub-$0.50/M-input models — DeepSeek V3.2, MiniMax M2.5, Qwen3-235B, Gemini 2.5 Flash Lite — with worked cost math.
OpenKey team
Free LLM APIs in July 2026: what's actually available
The 25 free models on OpenKey as of July 2026 — context windows, labs, and the honest caveats about free-tier rotation you should plan around.
OpenKey team
The real math of LLM API fees
How LLM gateway fees actually work, with real per-token math from Claude Sonnet 4.5 and a breakdown of what OpenRouter, Together, and Fireworks charge.
OpenKey team
Rankings you can't game: why we whale-filter
How one customer's traffic distorted a major LLM leaderboard, and the whale-filtering, promo-price flags, and published methodology we use to prevent it.
OpenKey team