🧠 Cheaper, faster, bigger
DeepSeek has unveiled V4.1 Flash, a 552-billion-parameter AI model that already outperformed the flagship V4 Pro in coding and agent benchmarks.
The company also cut memory demands by 4× and slashed off-peak API pricing by 50%:
↖️ https://durovscode.com/deepseek-v4-1-flash-ai-model-api-pricing
Post #4634
1.17K
- 👍 12