Blog · Category
LLM Waves — API pricing for every major model provider, GPT-6 Astra to DeepSeek-V4-Flash, compared on per-token rates, long-context multipliers, cache costs and the fees vendors don't put on the rate card.
5 articles
Perplexity retires Sonar on Sept 27, 2026. See Agent API preset pricing and how each preset's cost compares to the Sonar tier it replaces, with rate limits.
8 min read
Groq pricing for ten hosted models, checked against Groq's deprecation schedule. Two retired models still sell on Groq's pricing page as of Sept. 21, 2026.
10 min read
Mistral API pricing, September 2026: 8 active models from $0.10 to $7.50 per million tokens, plus 5 retired families most trackers still bill as current.
10 min read
Grok API pricing for every live model as of September 2026: token rates, the 200,000-token cliff, tool fees, rate limits and whether xAI has a free tier.
11 min read
OpenAI API cost, aggregated from OpenAI's own developer docs: GPT-6 Astra, Sol, Terra and Luna rates, the 272K token cliff, and the real cache write cost.
12 min read