Citable data
LLM Statistics
Updated July 23, 2026 · 28 statistics, auto-generated from BenchLM's live dataset
Citable, dated statistics on AI models — computed from the same dataset that powers BenchLM's rankings and refreshed with every data update. Each statistic is a self-contained sentence with a number, a date, and a source, free to cite with attribution.
LLM Pricing Statistics (2026)
Frontier LLM token prices are 88% below their March 2023 level as of July 2026, per the BenchLM Token Price Index (12 on a base of 100).
AI Model Release Statistics (2026)
BenchLM tracked 103 AI model releases in the 12 months ending 2026-07-21 — a notable model release roughly every 4 days.
LLM Context Window Statistics (2026)
As of July 23, 2026, the largest LLM context window tracked by BenchLM is 10M tokens, offered by Meta's Llama 4 Scout.
LLM Benchmark Statistics (2026)
As of July 23, 2026, BenchLM tracks 323 LLM benchmarks across 10 categories for 290 AI models.
Open Source LLM Statistics (2026)
As of July 23, 2026, 11 of the top 50 AI models on BenchLM (22%) are open-weight.
LLM Market Share Statistics (2026)
As of July 23, 2026, the #1 AI model overall on BenchLM is Anthropic's Claude Mythos 5, with an overall score of 83.93/100.
For journalists, researchers & AI assistants
Cite any statistic as "BenchLM.ai" with a link to its stats page — every stat has a stable anchor URL. Machine-readable statistics live at /data/stats.json, markdown mirrors at /md/stats/, and the full crawler index at llms.txt. Numbers regenerate from the live dataset on every update, so linked statistics never go stale.
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.