I still have a lot of results from DeepSeek-V4.1-Flash testing.
Here is a fireworks animation made by flash, compared to other models i have run, this one is the best. Again, three.js and one-shot, not through any harness. First frame is dark, so thumbnail is broken, play
Deepseek v4.1 flash is a great model (in most harness including hermes)
It's speed is wild BUT see how much it thinks for one simple request LMAO thank god the tokens are dirt cheap
Guys, imagine all this with one prompt via #deepseek and only one prompt
Astra-6 vs DeepSeek-v4.1 cloud vs local.
Same water prompt. Left → right: • Astra: “Worked for 30s” in ChatGPT • DeepSeek V4.1 Cloud: 103s, 28.5K completion tokens • DeepSeek V4.1 Local: 4× DGX Spark, 25m to HTML through Hermes 512×512 mesh. Click-to-ripple. Interacting
You have reached the end of the archive
All of deepseek