Codex users!! I tested GPT-6.1 Sol at max
And.... it is BAD I gave it a prompt to create a motion graphics video it used: - 2% weekly limit (15% 5 hr limit) - 1.7 Million tokens - 58 minutes of work - $1.8 API usage So... the model is EXTREMELY efficient and usage for
Inspired by karpathy's test last month, I decided to see how good Opus 5.5 was at rendering 3D graphics.
I'm taken aback by how good the sound design is more than the graphics themselves. I gave it the first few paragraphs of Dracula, and asked it to recreate the entire thing
Part 5: The Infinitely Breathing Cosmos
New video! Θ 14 ∞ to Θ 16c ∞ Homo Luminaris · part 3a · Ξ Ο Π Everything falls apart. Thirteen plates in one point. From there the Mandelbrot. Network. Grounding. Celebrate every little thing. 432.35 is correct. 34 ripples. Beating is 1. Mathematics
Anthropic has just released the official prompt word guide for Opus 5.5.
If you still write prompt words according to the Opus 5 method, you are probably wasting tokens. The changes are: 1. Default to medium intensity The Opus 5.5 already beats the high-intensity Opus 5 at medium intensity and costs less. 2. Stop writing "think step by step"
You have reached the end of the archive
All of Claude Opus 5