What Glevd does here
Glevd founded BenchLM and runs it: collecting benchmark results from their public sources, maintaining the scoring method, and writing the analysis on the blog. 110 articles carry this byline, the first from August 2025.
Rankings follow the published method, not the author's preference or a sponsor's. The editorial policy sets out how sourced facts, measured results and commercial relationships are kept apart, and how errors are corrected on the page.
Recent articles
- Claude Cowork has no price of its own. You pay for usage, November 20, 2026
- What a system prompt does, and where extra rules pay off, November 19, 2026
- The best AI chatbot, scored on what you can actually buy, November 17, 2026
- Cowork and chat merged, so choose Claude or Claude Code, November 13, 2026
- LLM temperature, and the models that no longer take it, November 12, 2026
- OpenAI Daybreak, explained: what Blue and Red each unlock, November 12, 2026
- ChatGPT or Claude for coding, by score and by cost, November 10, 2026
- How to evaluate an LLM before you pick one, November 10, 2026
- LLM News: Kimi K3.1 Pro Goes Open, GPT-6 Cyber Stays Gated, November 7, 2026
- GPT-6 models compared: Astra, Sol, Terra and Luna, October 29, 2026
Contact
Corrections and questions go to hello@benchlm.ai. The public account is @glevd on X.