Do you use Supermaven?
What is Supermaven?
Supermaven's new neural network architecture surpasses Transformers (the current standard architecture) in efficiency, integrating information across a 300,000-token context window at the cost and latency of a 4,000-token Transformer window.
Recent launches
Supermaven
Supermaven's new neural network architecture surpasses Transformers (the current standard architecture) in efficiency, integrating information across a 300,000-token context window at the cost and latency of a 4,000-token Transformer window.