I put DeepSeek-V4.1-Flash against GPT-6-Astra to see if either of them were capable of retopologizing a (horribly malformed) facial mesh via Blender.
The results were....not great. But it's a start! We should all pay close attention because these models are just going to get better and better at tasks like these.
This was my first experiment and both models completed their task in under an hour with very loose instructions. I suspect that with more specific prompting and a longer-running goal, we could expect a better result.
Interested in Blender add-ons to help with rigging and animation? Go to https://holotype.com.sg and https://nickfisher.gumroad.com/.
Dilarang memposting ulang tanpa izin dari Kreator.
I put DeepSeek-V4.1-Flash against GPT-6-Astra to see if either of them were capable of retopologizing a (horribly malformed) facial mesh via Blender.
The results were....not great. But it's a start! We should all pay close attention because these models are just going to get better and better at tasks like these.
This was my first experiment and both models completed their task in under an hour with very loose instructions. I suspect that with more specific prompting and a longer-running goal, we could expect a better result.
Interested in Blender add-ons to help with rigging and animation? Go to https://holotype.com.sg and https://nickfisher.gumroad.com/.