6 ms·

I really would like to see OpenAI’s focus on efficiency but everytime I use Codex, it wastes tokens like there’s no tomorrow, hitting week limit in a day, where

by rebolek 29d ago

I really would like to see OpenAI’s focus on efficiency but everytime I use Codex, it wastes tokens like there’s no tomorrow, hitting week limit in a day, where I’m able to use Claude just fine. Maybe it’s based on the codebase, I don’t know, but I have better results with Claude than Codex.

jimmydoe 29d ago
same boat. gpt 5.5/5.6 goes further on its own much more often than opus 4.8/5 does. codex capped ~300k context when claude does 1m. I don't feel codex is saving tokens, and result is usually not as good imo.
- gatio 29d ago
  > codex capped ~300k context when claude does 1m. That's configurable in codex.. but there is a higher cost/usage to using it.
bicepjai 29d ago
I had the opposite experience. Claude models and its harness feel like they are set to eat tokens for everything, especially if it’s ultracode effort. I have seen it spawn 6 agents and eat my 4-hour quota right in front of me. Codex is on point and follows instructions well even with max effort. I like my analogy of Claude being garrulous and Codex being laconic. If I see more limit hits in same sitting session, I have to rearrange my workflows. my experiments with qwen and deepseek have been good, cant wait to try glm and other models.
- throwup238 29d ago
  Just before I canceled my 20x Max sub I had a day where I had ~15% of my weekly I was trying to burn. I set it on ultracode and because I had set the max agents 32 for another project and forgot, the five research agents ended up spinning up a total of 26 subagents and burned through the remainder of my weekly in the span of 20 minutes before I noticed and shut it down. “No, not like THAT!”
  - vanviegen 29d ago
    How can you burn through your weekly budget in 20 minutes? You'd hit your 5 hour budget way before that, right?
    - throwup238 29d ago
      The five hour budget is about 15% of the weekly on Max 20x. I had about 15% left.
      - vanviegen 28d ago
        Ah sorry, I read your earlier comment as you being at 15% of your weekly (so the prompt consuming 85%).
  - wccrawford 29d ago
    Claude did the same thing to me the other day. I ran out of my 5 hr limit, and it went into my $100 credit they had given me. Multiple parallel agents. It ate it in no time flat. When I checked, all that credit was gone, I still wasn't into the next 5 hours, and all the agents had failed, returning nothing. I didn't even get anything for burning all that credit. If I had paid for it, I'd be very, very pissed.
    - newAccount2025 29d ago
      Similar experience here. I had raised my cap to $20 because they had given me promo credit. The credit expired without me using it and I didn’t notice. One fine day I was like, oh my weekly quota resets soon, I will kick off an expensive bug hunt. It launched parallel things and burned $10 in like a moment. It’s scary. I turned off extra usage after that.
- StilesCrisis 29d ago
  Requesting ultracode is basically asking for maximum token usage. If you want to limit consumption, use medium or high.
- NyxWulf 29d ago
  Here is an excerpt from the system prompt for UltraCode (Same for Fable,Opus,Sonnet): "Ultracode. When a system-reminder confirms ultracode is on, that opt-in is standing: author and run a workflow for every substantive task by default. The goal is the most exhaustive, correct answer you can produce — token cost is not a constraint."
- killix 28d ago
  [flagged]
vorticalbox 29d ago
It might be your prompting. My boss uses opus and gets good results when I use it always burns tokens. The other way is also true too I get great results with sol/terra but my boss does not.
- rebolek 29d ago
  Maybe. But if Claude is fine with my prompts while Codex uses them as an excuse to burn tokens. I have nothing against the model, they are more or less on pair, it’s just that Claude is giving me more value for same money.
  - vorticalbox 29d ago
    If Claude is working for you then stick with it! Happy you’re getting good value from Claude.
- ygjb 28d ago
  One of my challenges is that I often end up burning tokens trying to build the right solution rather than building a solution that is good enough and deferring the right choices until later. I have been actively working to change my expectations for building with AI to accept worse solutions to get things going rather than trying to solve everything at launch. This has been a recurring problem for me - as a security engineer catasrophization is a fundamental skill to finding vulnerabilities in complex systems, but makes me to conservative when building. For some of my leaders they are much better at saying 'good enough, ship it, and fix it later'.
llama052 29d ago
At least with OpenAI you can use a third party harness like Pi.
shepherdjerred 29d ago
Are you kidding me? OpenAI weekly limits feel _infinite_. I run out of Anthropic limits reliably halfway through the week
chaboud 29d ago
Codex is fabulous at work, where token use is near limitless and ultra-thorough tool use is welcome. Go ahead and fire off searches for look-alike terms on my 96-core cloud instance. By contrast, Claude Code's bias to make assumptions of reasonableness about underlying systems has proven to be immensely frustrating over the last month or two, both personally and at work. I've wasted days on "that was my mistake. I've been reporting numbers on the old architecture because I hadn't enabled the new one in the config" both at work and home. It's immensely frustrating. But here we are. Wrestling with energetic idiots in model form, wrangled by over-specific harnesses that struggle to stay off of deranged side-quests. What a time to be alive!

by rebolek 29d ago

jimmydoe 29d ago

same boat. gpt 5.5/5.6 goes further on its own much more often than opus 4.8/5 does. codex capped ~300k context when claude does 1m. I don't feel codex is saving tokens, and result is usually not as good imo.

gatio 29d ago
> codex capped ~300k context when claude does 1m. That's configurable in codex.. but there is a higher cost/usage to using it.

bicepjai 29d ago

I had the opposite experience. Claude models and its harness feel like they are set to eat tokens for everything, especially if it’s ultracode effort. I have seen it spawn 6 agents and eat my 4-hour quota right in front of me. Codex is on point and follows instructions well even with max effort. I like my analogy of Claude being garrulous and Codex being laconic. If I see more limit hits in same sitting session, I have to rearrange my workflows. my experiments with qwen and deepseek have been good, cant wait to try glm and other models.

throwup238 29d ago
Just before I canceled my 20x Max sub I had a day where I had ~15% of my weekly I was trying to burn. I set it on ultracode and because I had set the max agents 32 for another project and forgot, the five research agents ended up spinning up a total of 26 subagents and burned through the remainder of my weekly in the span of 20 minutes before I noticed and shut it down. “No, not like THAT!”
- vanviegen 29d ago
  How can you burn through your weekly budget in 20 minutes? You'd hit your 5 hour budget way before that, right?
  - throwup238 29d ago
    The five hour budget is about 15% of the weekly on Max 20x. I had about 15% left.
    - vanviegen 28d ago
      Ah sorry, I read your earlier comment as you being at 15% of your weekly (so the prompt consuming 85%).
- wccrawford 29d ago
  Claude did the same thing to me the other day. I ran out of my 5 hr limit, and it went into my $100 credit they had given me. Multiple parallel agents. It ate it in no time flat. When I checked, all that credit was gone, I still wasn't into the next 5 hours, and all the agents had failed, returning nothing. I didn't even get anything for burning all that credit. If I had paid for it, I'd be very, very pissed.
  - newAccount2025 29d ago
    Similar experience here. I had raised my cap to $20 because they had given me promo credit. The credit expired without me using it and I didn’t notice. One fine day I was like, oh my weekly quota resets soon, I will kick off an expensive bug hunt. It launched parallel things and burned $10 in like a moment. It’s scary. I turned off extra usage after that.
StilesCrisis 29d ago
Requesting ultracode is basically asking for maximum token usage. If you want to limit consumption, use medium or high.
NyxWulf 29d ago
Here is an excerpt from the system prompt for UltraCode (Same for Fable,Opus,Sonnet): "Ultracode. When a system-reminder confirms ultracode is on, that opt-in is standing: author and run a workflow for every substantive task by default. The goal is the most exhaustive, correct answer you can produce — token cost is not a constraint."
killix 28d ago
[flagged]

vorticalbox 29d ago

It might be your prompting. My boss uses opus and gets good results when I use it always burns tokens. The other way is also true too I get great results with sol/terra but my boss does not.

rebolek 29d ago
Maybe. But if Claude is fine with my prompts while Codex uses them as an excuse to burn tokens. I have nothing against the model, they are more or less on pair, it’s just that Claude is giving me more value for same money.
- vorticalbox 29d ago
  If Claude is working for you then stick with it! Happy you’re getting good value from Claude.
ygjb 28d ago
One of my challenges is that I often end up burning tokens trying to build the right solution rather than building a solution that is good enough and deferring the right choices until later. I have been actively working to change my expectations for building with AI to accept worse solutions to get things going rather than trying to solve everything at launch. This has been a recurring problem for me - as a security engineer catasrophization is a fundamental skill to finding vulnerabilities in complex systems, but makes me to conservative when building. For some of my leaders they are much better at saying 'good enough, ship it, and fix it later'.

llama052 29d ago

At least with OpenAI you can use a third party harness like Pi.

shepherdjerred 29d ago

Are you kidding me? OpenAI weekly limits feel _infinite_. I run out of Anthropic limits reliably halfway through the week

chaboud 29d ago

Codex is fabulous at work, where token use is near limitless and ultra-thorough tool use is welcome. Go ahead and fire off searches for look-alike terms on my 96-core cloud instance. By contrast, Claude Code's bias to make assumptions of reasonableness about underlying systems has proven to be immensely frustrating over the last month or two, both personally and at work. I've wasted days on "that was my mistake. I've been reporting numbers on the old architecture because I hadn't enabled the new one in the config" both at work and home. It's immensely frustrating. But here we are. Wrestling with energetic idiots in model form, wrangled by over-specific harnesses that struggle to stay off of deranged side-quests. What a time to be alive!