An AI agent demo shows GPT-5.6 Sol executing rapid 3D modeling inside Blender 4.5.4 at 750 tokens per second
The system used GUI controls instead of Blender's Python API.
Combined views
1.1M
11 Sources, first seen ago
Sources
CH 
Chubby♨️@kimmonismusGPT-5.6 Sol on 750 Token/s in Blender. Not sped up. Holy moly. That’s so much more impressive than looking at benchmarks https://x.com/evayzh/status/2075374512401711203/video/1
- likes: 4.2K
- replies: 146
- bookmarks: 1.4K
- reposts: 228
CM Charlie Marsh@charliermarsh5.6 Sol Ultra on fast mode lol https://twitter.com/ChrissGPT/status/2075392201564328263
- likes: 25
- replies: 2
- bookmarks: 2
- reposts: 0
TI Tibo@thsottiauxThis actually not the version of GPT-5.6 Sol on Cerebras. Just normal /fast mode. Running it at 750+ TPS is a whole other thing. https://twitter.com/kimmonismus/status/2075482486901969066
- likes: 2.5K
- replies: 171
- bookmarks: 316
- reposts: 98
GB Greg Brockman@gdbComputer use with SOL Ultra: https://twitter.com/chrissgpt/status/2075392201564328263
- likes: 492
- replies: 32
- bookmarks: 63
- reposts: 29
ND Nick Dobos@NickADobosThe way codex computer use works is super cool In this scenario it’s not doing each computer use step one at time, Sol generates code on the fly which calls clicks and typing events It’s writing tiny throwaway scripts. That’s why it’s so fast, the AI isn’t actually doing it…
- likes: 163
- replies: 14
- bookmarks: 52
- reposts: 8