semisol on Nostr: It seems that part of GPT 5.6’s token efficiency came from reducing the required ...
It seems that part of GPT 5.6’s token efficiency came from reducing the required “thinking budget” for high/xhigh.
So, xhigh is no longer truly xhigh.
Published at
2026-07-11 05:59:46 UTCEvent JSON
{
"id": "ca2a3bfbaaedb5decbc39c8c7e540249edd7288346e92c6cb8615c6f9f449556",
"pubkey": "52b4a076bcbbbdc3a1aefa3735816cf74993b1b8db202b01c883c58be7fad8bd",
"created_at": 1783749586,
"kind": 1,
"tags": [
[
"client",
"Damus"
]
],
"content": "It seems that part of GPT 5.6’s token efficiency came from reducing the required “thinking budget” for high/xhigh.\n\nSo, xhigh is no longer truly xhigh.",
"sig": "f971b7307483a0e1c6000e019bf893c26b9c9912ba0d9c1f6f19a234ae19b0e1b099ba4263e25a96dc1aa08cfe1295652abb26394c61dfc2368609b90e7bc4f8"
}
