{"component":{"type":17,"accent_color":7742708,"components":[{"type":9,"components":[{"type":10,"content":"## [OpenHermes 2.5 Mistral 7B](https://openrouter.ai/teknium/openhermes-2.5-mistral-7b)\nteknium · 4K context"},{"type":10,"content":"A continuation of OpenHermes 2 model, trained on additional code datasets. Potentially the most interesting finding from training on a good ratio (est. of around 7-14% of the total dataset) of code instruction was that it has boosted several non-code benchmarks, including…"}],"accessory":{"type":2,"style":5,"label":"Open","url":"https://openrouter.ai/teknium/openhermes-2.5-mistral-7b"}},{"type":14,"divider":true,"spacing":1},{"type":1,"components":[{"type":2,"style":5,"label":"Chat","url":"https://openrouter.ai/chat?models=teknium%2Fopenhermes-2.5-mistral-7b","emoji":{"name":"💬"}},{"type":2,"style":5,"label":"Compare","url":"https://openrouter.ai/compare/teknium/openhermes-2.5-mistral-7b"},{"type":2,"style":5,"label":"API","url":"https://openrouter.ai/teknium/openhermes-2.5-mistral-7b/api"}]}]}}