DeepSeek is getting stupider and stupider. Do you still dare to use such a model?
DeepSeek releases a new series of models: deepseek-v4.1-flash-expires-on-0910
Claimed to: Adopt a new model structure, native multi-modal support, stronger capabilities, faster speed, and lower cost. The first time I put it on Pi Agent and ran it, I felt that 300 token/s was too little, just one word, quick Usage steps: Directly put deepseek-v4-flash…
DeepSeek V4.1 Flash just dropped, and holy shit this thing is fast,I’m getting almost 400 t/s in my video tests.
The speed is honestly crazy, I’m running my own benchmark right now and I’ll share all the results once I’m done. You just need to go to the DeepSeek harness and use
Deepseek v4.1 internal beta version released
Just change your model name to deepseek-v4.1-flash-expires-on-0910 and you can call it in the official API! The price is the same as deepseek-v4-flash, native multi-modal And because not many people use it, it’s so fast now, it can reach 400tok/s.
You have reached the end of the archive
All of deepseek