Granite 4.0 H Micro
Neural Network
Granite 4.0 H Micro is a compact 3B parameter IBM language model, available via API for integration into your products.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 131 000 tokens
- Max answer length
- 117 900 tokens
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Granite 4.0 H Micro?
Granite 4.0 H Micro is a compact model from IBM's Granite 4.0 family: 3 billion parameters make it lightweight and fast where a large model is overkill. It is designed for everyday text tasks: analyzing documents and correspondence, generating and editing code, drafting, structuring notes, and answering based on provided material. Via BotHub, the model is accessible without a VPN or foreign card; payment is made with a Russian card in rubles, only for used tokens, with no subscription, and tokens do not expire. With over 250 models available in one window, it's easy to compare Granite's responses with others and switch to the right one; a unified OpenAI-compatible API allows this without rewriting integrations. Data is transmitted with AES-GCM encryption, and companies have access to contracts, invoices, electronic document management (EDO), and an admin panel with limits. Developers use this model as a fast layer for code generation and analysis, support teams for classifying tickets and drafting responses, analysts for summarizing long documents, and product teams for assistant prototypes where response speed is critical.Frequently asked questions about Granite 4.0 H Micro
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.
Granite-4.0-h-micro is a compact text model from the Granite 4.0 line. It answers questions, writes and edits texts, analyzes documents, helps with code, and analyzes images. Via BotHub, the model is available from Russia without a VPN or foreign card, payment is pay-as-you-go, and tokens do not expire.
In a single response, the model outputs up to 117,900 tokens, which is a very large volume of text. The context window reaches up to 131,000 tokens, so both voluminous source material and the finished result fit in one window. This is enough for long documents and large data files.
There is no specific note about a reasoning mode before answering in our data, so we will not promise step-by-step reasoning. The model from the Granite 4.0 line answers directly. If a long logical chain is important for your task, compare it with another model in BotHub using your own examples.
There is no note about function calling and structured JSON output in our data, so we cannot confirm this for sure. Check the model documentation or test it with your request. If your system requires strict JSON, there are models with confirmed support nearby in BotHub.
The model accepts images and documents as input, as stated in its capabilities. Audio and video are not declared in the description, so you should not upload them. Text, images, and documents can be conveniently sent via BotHub, which is accessible from Russia without a VPN or foreign card.
We do not have specific data on the quality of Russian language, so we will not promise perfect results. The model confidently handles general text tasks, but Russian responses should be checked with your own examples. In BotHub, you can switch to another model and compare the result on the same request.