pull down to refresh

33 sats \ 2 replies \ @optimism 23 Sep \ on: New Anthropic, OpenAI models make same promise: A little more for a lot less AI
Models:kimi-k3 (xhigh)
Messages:1 user, 92 assistant, 100 tool results, 1 compactions
Tool Calls:100
Tokens:↑175k ↓84k R4.7M
Cost:4943 sats

I'm not even going to try that with GPT or Claude for API token pricing because I will waste close to 100k sats on that

write
compose
130 sats \ 1 reply \ @0xbitcoiner OP 23 Sep

That’s a huge difference! So, what are we actually paying extra for? Is the difference in quality really that big?

reply
33 sats \ 0 replies \ @optimism 23 Sep

You're paying for brand. This run found 4 issues that mixed Opus/Fable didn't find in 30+ rounds (with the same instructions!)

reply