Wow. That settles it. I'm also getting same API eq usage ($1500-1600~) but you know the funny part? I'M ON 20X PLAN. I think I'm being scammed lol. @trq212@bcherny@lydiahallie WHAT THE FUCK GUYS?? Any explanation? Any help? Wtf am i supposed to do when my acc is glitched likets
Here you go! Ran ccusage for this week and filtered out Codex.
Claude only:
• $1,420.61 API-equivalent usage
• 408M cache reads
• 12.26M output tokens
• 6.24M cache writes
• Mostly Opus 5.5
And my weekly usage meter was at 82%.
These are local usage estimates, not actual charges. Curious how yours compares!
@sammcallister I've been robbed of my week limits on my 20x plan being insultingly close to a 5x plan. For a month straight. I've now downgraded to 5x and you guys lost my permanent trust.
@apples_jimmy And the agent in chatgpt currently can not specify images to the image model, it can only give the prompt and the relevant images are selected by the backend, this pollutes the reference image list used to generate the image and can exclude images that wereneeded to be included.
@apples_jimmy Hey idk if u have contact with openai, but can you please tell them to specify which image model name and version was used to generate an image in its provenance/manifest metadata?
And to let chatgpt/codex be able to specify image resolution to the imagen model.
@finnardley I don't think you can match their quality in terms of animation quality density and information density and the narration and the flow and the way they subtly optimize the video for more watchtime.
@theo@dominikkoch It still has UX issues. Maybe the Send button could turn yellow when Compact is enabled, to signal that it performs a different action. I’ve accidentally sent messages with Compact on multiple times without realizing it.