llm-api
AIDiffusion LLMs explained: what Mercury 3's 1,000 tokens per second means for builders
Inception's Mercury 3 is a diffusion-based language model running at 1,068 tokens per second with a 1M-token context. Here's how diffusion LLMs differ from ChatGPT-style models, what Mercury 3 costs, and where speed matters.
AIGemini 4 Argon for builders: pricing, caching, and when to pick it
Gemini 4 Argon launched on Oct 13 at an introductory $2/$10 per million tokens, with cached input 95% off. The price doubles after Nov 30. Here's what builders get, where it runs, and how to decide.
AIGPT-6 explained: Sol vs Luna vs Astra, and what \"Intelligent UI\" means
GPT-6 is a family, not one model. Astra is the top model and the API, Sol runs paid ChatGPT, Luna runs Free and Go. Here's which one you're using, what Astra costs, and what OpenAI's benchmarks actually cover.
AIMiniMax M3.1 Flash: a 73.7% DeepSWE coding model at $0.15 per million tokens
MiniMax M3.1 Flash, the model that ran anonymously as \"Space Bunny\", costs $0.15/$0.60 per million tokens and scores 73.7% on DeepSWE v1.1 by MiniMax's count. The open weights are now on Hugging Face under the MiniMax Community License. Here's what's confirmed and what the license allows.
Build GuidesPrompt caching explained: how to cut your AI API bill without changing models
Prompt caching lets you pay a fraction of the price for input the model has seen before. Here's how it works, what it costs on Claude, Gemini, GPT-6 and MiniMax, and how to structure prompts so the cache actually hits.