I've found that using Claude (in Terminal.app) with DeepSeek at the back-end is much the same (I'd say about 95% of Real-Claude's ability), and it's a lot cheaper. I can burn through the 20x weekly limit pretty much every week (and not have credit for the weekend) so that's costing me £200/month. DeepSeek is averaging at about £1/day.
I just bumped my Claude account down from the £200/month option to £18/month. I use Real-Claude to validate the decisions of DeepSeek, which seems to not get quite as much of the fine details as Real-Claude does. Once they're pointed out though, DeepSeek is fine, and off to the races.
They're trying to win back the people they pushed away by only increasing the weekly limit by 50%. I'm like, "What?"
Won't go back, especially since their new Opus 4.7 is such a disaster.
I've been using opus 4.7 on xhigh effort and have been blown away by the results
done the same and I received email from corporate that I'm not allowed to do that :) But it's been great while using it. The only downside was the long wait time for each response.
Why are you not allowed to do that?
I'm just hoping they release Mythos soon now that it seems like they have enough compute to do promotions like this
I really doubt that will happen anytime soon. I imagine there is still a great deal of preventative security-related work going on right now before they'll consider releasing this to the public (especially as this means granting access to Mythos to malicious actors).
Great news for all developers at the Anthropic casino.