Hyperstack Hosted Models

Open-weight base models hosted on Hyperstack for inference and fine-tuning.

Hyperstack hosts a set of open-weight base models directly on its own infrastructure, so you can run inference and, for supported models, fine-tuning without relying on an external provider. This page introduces the base models available on Hyperstack and the maximum context length that each one supports.

Available Models

The table below summarizes the base models that are hosted and served directly on Hyperstack infrastructure, along with descriptions and links to their official documentation and Hugging Face model pages.

Model Name	About the Model
gpt-oss-120b	A powerful 120B parameter model released by OpenAI as its first open-weight model since GPT-2. Designed for advanced reasoning and tool use, it performs well on instruction-following, long-context understanding, and math tasks. Ideal for building high-performance generative applications in research or production environments. See documentation See Hugging Face model page Model name for API: openai/gpt-oss-120b Note: This model supports inference only and does not yet support fine-tuning.
Mistral Small 24B Instruct 2501	A compact 24B model designed for low-latency inference and resource-constrained environments. Despite its size, it delivers strong results in instruction following, math, and code generation tasks. Ideal for real-time applications like chatbots, support agents, and embedded systems. See documentation See Hugging Face model Model name for API: mistralai/Mistral-Small-24B-Instruct-2501
Llama 3.3 70B Instruct	Meta’s flagship 70B parameter model, fine-tuned for instruction-based tasks. It is optimized for complex reasoning, tool use, and long-context generation (up to 128K tokens). Suitable for production use in advanced multilingual chatbots, coding assistants, and text analysis systems. See documentation See Hugging Face model Model name for API: meta-llama/Llama-3.3-70B-Instruct
Llama 3.1 8B Instruct	A smaller alternative to the 70B variant, this 8B model provides a good balance of performance and efficiency. It supports a broad range of general-purpose tasks while being more cost-effective for development and experimentation. See documentation See Hugging Face model Model name for API: meta-llama/Llama-3.1-8B-Instruct

Context Length Limits

Context length refers to the total number of tokens that can be processed in a single interaction using the Chat Completions API or the Playground UI in AI Studio. This includes both the input (system prompt, prior messages, and user prompt) and the model’s generated output.

The maximum supported context length varies by model. See the table below for details:

Model Name	Max Context Length
openai/gpt-oss-120b	128k tokens
mistralai/Mistral-Small-24B-Instruct-2501 meta-llama/Llama-3.3-70B-Instruct meta-llama/Llama-3.1-8B-Instruct	8192 tokens

Back to top