Morph V3 Large
Neural Network
Morph V3 Large is an apply-model for code: makes complex edits with 98% accuracy at ~4500 tokens/sec.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Morph V3 Large?
Morph V3 Large is a model by the Morph team designed for one task: accurately applying edits to existing code. It accepts instructions and source files in a special format with 'instruction' and 'code' tags and returns the modified code. With 98% accuracy and a speed of ~4500 tokens/sec, it's ideal for the end of a chain where another model generates changes, and Morph V3 Large applies them. In BotHub, it's accessible from Russia without a VPN or foreign card; pay in rubles with Russian cards only for tokens used—they don't expire. A unified OpenAI-compatible API lets you connect the model to your editor or CI and switch to any of 250+ other models without rewriting the integration. Useful for applying diffs to large files, mass refactoring, implementing reviewer feedback, automating linter fixes, or building a custom code agent where you handle the plan and the model handles precise editing.Frequently asked questions about Morph V3 Large
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
morph-v3-large is a text model from Morph focused on code and fast file editing. It accepts text, documents, and images as input, with a 32,000-token context window. Suitable for refactoring, debugging, precise snippet editing, and parsing technical documentation.
The output limit is 131,072 tokens, which is substantial: the model can return a large file in full or multiple edited snippets at once. The request context is 32,000 tokens, so it's better to provide large source files in parts.
A dedicated reasoning mode is not specified in our data. However, you can ask the model in the prompt to break down the task step-by-step and explain the editing logic—this is usually sufficient for coding. If you need long chains of thought, switch to a reasoning model in the same window.
Function calling and strict JSON mode are not specified in our data, so you should test this for your specific agent scenarios. You can always describe the required output format in the prompt. BotHub's unified OpenAI-compatible API allows you to switch to another model without rewriting your integration.
The model accepts images and documents along with text: attach an error screenshot, diagram, or spec file. Audio and video are not supported. The model only generates text, so you won't get images, audio, or video from it.
We don't have specific data on language support, so we can't make promises—it's best to test the model with your real request. The BotHub interface is in Russian, access works from Russia without a VPN or foreign card, and payment is in rubles based on actual usage.