Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Haiku, a new model, priced at $0.25/1M tokens, is generating excitement in the coding community as it competes with GPT-4. Users believe Haiku could surpass GPT-3.5 in coding capabilities and offer a cheaper alternative with 200k context. Haiku is praised for rapid prototyping and code completion potential.
Haiku is a pretty decent model for code and could be a strong candidate for a code completion model (kind of like what GPT-3.5 is in Copilot/Cursor).
Claude Haiku is the best for rapid prototyping as it is super fast and smart. After lots of iterations, Haiku is exploring the world of Dune and expressing it visually in three.js: "Imagine a planet where the very sands beneath your feet hold the key to the future of humanity,… https://t.co/2Q8V1dAcJB
I just wanted to say that theoretically Haikku is a better coder than Sonett.. If this is real and replicable in real life we have a better model than Gemini Ultra 1.0, GPT-4 (Non-turbo) in Code for a cheaper price than GPT-3.5 and with 200k context. How many things this change?