I’ve been using GPT-5.6 ☀️ in my daily work, and it’s exceptionally good at tracing deep, complex bugs back to their actual root cause not just patching the symptoms.
@kimmonismus Sol high designs
↓
Sol medium implements substantive work
↓
Terra high reviews independently
↓
Sol high integrates and verifies
works fine on my end, it stops spawnin 20 Subs.
@realhasanshoaib@thegenioo The important surprise:Luna xhigh is only about 3% cheaper than Terra high and 19% cheaper than Sol medium in Artificial Analysis complete evaluation not dramatically cheaper. Its low token price is offset by enormous output volume Sol medium: 12M Terra high: 24M Luna xhigh: 67M
@kimmonismus Luna xhigh is only about 3% cheaper than Terra high and 19% cheaper than Sol medium in Artificial Analysis complete evaluation not dramatically cheaper. Its low token price is offset by enormous output volume
Sol medium: 12M Terra high: 24M Luna xhigh: 67M
@kimmonismus Sol medium is okaish, but I let Terra high handle diff review and use Luna as the explorer. However, Sol high and above is absolutely unusable!