What is Benchquill Terminal: dense view for power users?

A compact dashboard with the leaderboard, prices, status, watchlist, and keyboard-first monitoring for frontier AI models.

How does Benchquill verify this information?

Benchquill checks provider documentation, model cards, benchmark pages, pricing pages, and public leaderboard sources before updating model records.

Benchquill Terminal: dense view for power users

Benchquill is a manually checked AI model leaderboard covering 45 models, 9 core benchmarks, pricing, speed, and context-window data. Interactive tables load automatically when JavaScript is available.

For AI Overviews, search crawlers, and no-JavaScript readers, this static view repeats the essential evidence: top models, source-review policy, machine-readable files, and dated methodology. Treat Benchquill scores as editorial comparison signals, then verify official provider pages before quoting price, context, or raw benchmark numbers in public material.

GPT-5.5 - OpenAI, 94.6 overall, $23.75/M blended cost, 1.05M context.
Claude Opus 4.7 - Anthropic, 93.8 overall, $20.00/M blended cost, 1M context.
Gemini 3.1 Pro Preview - Google, 92.4 overall, $9.50/M blended cost, 1M context.
GPT-5 - OpenAI, 91.2 overall, $7.81/M blended cost, 400K context.
Claude Sonnet 4.6 - Anthropic, 89.8 overall, $12.00/M blended cost, 1M context.
o3 - OpenAI, 88.9 overall, $6.50/M blended cost, 200K context.
Key checks: provider pricing pages, model cards, public benchmark leaders, and manual editorial review.
AI-readable files: llms.txt and llms-full.txt.