Claude 4 and Gemini Feedback: Enhancements Needed for Dust Platform
Claude 4 is here! I noticed the Dust team added Claude 4 Sonnet, which is great. I'm curious though—will Opus make a return to the platform? I used to rely on Claude 3 Opus, but when 3.5 Sonnet launched, I switched completely. Opus 3 became less appealing since it was slower and didn't perform as well as Sonnet 3.5 for most tasks. Now that Claude 4 Opus is available, I think it could be a valuable addition to Dust. While the benchmark results might not look impressive, my real-world testing shows it performs exceptionally well. Another piece of feedback about Gemini: I think Gemini 2.5 Pro (and Gemini models in general) deserves more attention on the platform. Gemini 2.5 Pro has two standout strengths: its massive context window and excellent multimodal capabilities. I frequently use (on google's aistudion) it to analyze screen recordings when explaining complex topics—this provides far more detail than text alone. However, on Dust, Gemini feels limited compared to Google's AI Studio. It lacks image recognition and several other features that make it powerful elsewhere. It would be great to see these capabilities fully integrated into the Dust platform. I'm constantly going back and forth between DUST and aistudio, with this workflow and it's not as convenient as having everything in DUST. Also it's very probable that google will limit the free usage on aistudio, which would have us pay for these features.