Comment by LorenDB
12 hours ago
I've heard that DeepSeek v4 Flash 0731 has frequently assumed that it has vision capabilities and then resorts to inventing text-based image analysis tools when it finds that it actually can't see. In that case, this is a great upgrade for the model.
Anecdotally, I had to tell 0731 to refrain from viewing screenshots since it kept breaking its sessions by trying to read images.
It tried to recreate vision by analyzing pixels on 3 separate projects I had it working on.
I've mitigated this by giving it a "skill" that just means the harness using a different model.
Yeah I've seen it a lot. It goes through the effort, unasked, of pulling screenshots off a connected device and then it's like... Oh shit yeah I can't see.
It's doing it's best to accomplish whatever task you've thrown at it.
It's expecting you to have done at least something besides select DS4 on Ollama, essentially.
Even with the price hike, Deepseek V4 Flash still does this a lot better than any similarly priced model, in my experience. I've had Luna take shortcuts (like adding an overload to methods whose signature it changed so they don't break existing tests, instead of fixing the tests) or just not do the entire work and report it as done (did not fully resolve rebase conflicts). Deepseek has never really failed in this type of way for me, and it has been far more persistent in validating its work than Luna (and several bigger models).