
Pinkstack/Superthoughts-mini-v1-3.8b
Text Generation
•
Updated
•
78
•
•
1
Trained to reason properly.
Note 3.8B reasoning model based on phi-3.5, 128K context and significantly better than superthoughts lite.
Note Non-experimental version of lite. 1.7B parameters, GRPO & SFT trained, not very conversational though.
Chat with a lite reasoning AI with a web search tool online.
Note Free online chat with superthoughts lite, keep in mind that it is not very conversational, this space includes web-search abilities.
Note safetensors 14B
Note Non-experimental version of lite. 1.7B parameters, GRPO & SFT trained. GGUF version.
Note official gguf quants 14B
Note 3.8B gguf