an LLM as capable as GPT-4 runs on one 4090 by March 2025
11
99
1k
2025
27%
chance

e.g. Winograde >= 87.5%

Get Ṁ600 play money
Sort by:

does it count that I can run the llm while also using cpu ram offloading just like ollama does automatically? (it would be very slow, but would work)