@AcuWeb3@lidageai Most people have only heard of it because international news coverage inevitably mentions DeepSeek. In reality, hardly anyone actually uses it. Its API is technically cheap, but DeepSeek doesn’t offer heavily subsidized subscription plans like GPT and Claude do.
@Priyannkaaaa All I run is Opus 5.5 medium + sub Sonnet 5.5 high, plus Fable as advisor. Easily enough to comfortably cruise through the 5h session on my 100$ plan.
@cognition If we could use our ChatGPT sub (and hopefully Claude later) with Fusion mode, Devin would be an absolute game changer.
Just focus on the harness and Devin’s own model, then let users plug in top-tier models through their existing subscriptions.
@dabit3 Could there be a discount voucher for the Max plan, so people interested in Devin also get a chance? Not everyone is lucky enough to get it for free.
@dabit3 Fusion is super efficient. I tried Devin thanks to the SWE2 promotion, but I ended up preferring Fusion. However, Devin needs to reconsider its subscription pricing: going from $20 to $200 is too big of a gap. They should offer a midtier plan around $100
@NPC35982722@dummerspast39 Sorry if I came off a bit harsh earlier. I just hope the project won’t affect ChatWeb usage in the future. I don’t want it to get patched like what happened with Goal abuse when Codex quota was almost used up before.