Simon Willison’s Blog
Sponsor: Datadog — deliver reliable AI faster with LLM observability. Read the best practices guide
May 20, 2026 - link blog
How fast is 10 tokens per second, really? (source) Mike Veerman built a neat little HTML app (source code here) that lets you simulate LLM token generation speeds, ranging from 5 tokens per second up to 800 tokens per second.
If you see a model advertised as “30 tokens per second” and want an intuitive feel for what that actually looks like, this tool is useful.
Posted on May 20, 2026 at 5:57 PM
Recent articles
- Gemini 3.5 Flash: more expensive, but Google plans to use it everywhere - May 19, 2026
- The last six months of LLMs: a five-minute tour - May 19, 2026
- Notes on the xAI/Anthropic data center deal - May 7, 2026
This is a link post by Simon Willison, published on May 20, 2026.
Monthly newsletter
Sponsor me for $10 a month and receive a carefully curated email digest covering the most important LLM developments of the month.
Pay me to email you less!