OpenChat_8192 Beats ChatGPT-3.5 on Vicuna GPT-4 Benchmark

The model comes in three versions: the basic OpenChat model, OpenChat-8192 and OpenCoderPlus (Code generation: 102.5% ChatGPT)<p>- OpenChat: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;openchat - OpenChat_8192: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;openchat_8192 (best chat) - OpenCoderPlus: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;opencoderplus (best coder)<p>- Dataset: https:&#x2F;&#x2F;huggingface.co&#x2F;datasets&#x2F;openchat&#x2F;openchat_sharegpt4_dataset<p>- Code: https:&#x2F;&#x2F;github.com&#x2F;imoneoi&#x2F;openchat The model comes in three versions: the basic OpenChat model, OpenChat-8192 and OpenCoderPlus (Code generation: 102.5% ChatGPT)<p>This is a significant achievement considering that it&#x27;s the first (released) open-source model to surpass the Vicuna benchmark.<p>- OpenChat: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;openchat - OpenChat_8192: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;openchat_8192 (best chat) - OpenCoderPlus: https:&#x2F;&#x2F;huggingface.co&#x2F;openchat&#x2F;opencoderplus (best coder)<p>- Dataset: https:&#x2F;&#x2F;huggingface.co&#x2F;datasets&#x2F;openchat&#x2F;openchat_sharegpt4_dataset<p>- Code: https:&#x2F;&#x2F;github.com&#x2F;imoneoi&#x2F;openchat