---
title: "Heroku AI Expands Model Offering with OpenAI’s gpt-oss-120b"
description: "Heroku has made OpenAI’s open-weight gpt-oss-120b available through Heroku Managed Inference and Agents, giving developers an OpenAI-compatible chat completions API for compatible SDKs and frameworks. The model uses a Mixture-of-Experts architecture with 117 billion total parameters, activating 5.1 billion per token and enabling operation on a single 80GB GPU; its Apache 2.0 license permits inspection and fine-tuning. It is designed for agentic workflows involving instruction following, function calling, web searches, and Python execution, while reported benchmarks place it near or above several open-weight models and sometimes OpenAI o4-mini. The add-on provides deployment, secure application access, agent actions, monitoring, and scaling without infrastructure management, with pricing of $0.15 per million input tokens and $0.60 per million output tokens."
---

# Heroku AI Expands Model Offering with OpenAI’s gpt-oss-120b

[Heroku](https://yomu.fyi/company/heroku) · Anush DSouza · Aug 20, 2025

**Type:** Announcement

## Summary

Heroku has made OpenAI’s open-weight gpt-oss-120b available through Heroku Managed Inference and Agents, giving developers an OpenAI-compatible chat completions API for compatible SDKs and frameworks. The model uses a Mixture-of-Experts architecture with 117 billion total parameters, activating 5.1 billion per token and enabling operation on a single 80GB GPU; its Apache 2.0 license permits inspection and fine-tuning. It is designed for agentic workflows involving instruction following, function calling, web searches, and Python execution, while reported benchmarks place it near or above several open-weight models and sometimes OpenAI o4-mini. The add-on provides deployment, secure application access, agent actions, monitoring, and scaling without infrastructure management, with pricing of $0.15 per million input tokens and $0.60 per million output tokens.

## Context

OpenAI released gpt-oss-120b as an open-weight model intended to make powerful AI more accessible to developers and enterprise teams. The model targets text generation, understanding, and agentic workflows, with reported benchmark performance competitive with proprietary and other open-weight models.

## Approach / What changed

Heroku offers gpt-oss-120b through its Managed Inference and Agents add-on, accessible via an OpenAI-compatible chat completions API. The platform handles deployment, secure access, agent execution, monitoring, and scaling, removing the need for teams to provision and manage inference infrastructure.

## Takeaways

- gpt-oss-120b has 117 billion total parameters but activates 5.1 billion per token through its Mixture-of-Experts architecture, allowing it to run on a single 80GB GPU.
- The model’s reported benchmarks include 90.0% on MMLU, 97.9% on AIME 2025 with tools, and a 2463 Elo rating on Codeforces without tools.
- Heroku lists pricing at $0.15 per million input tokens and $0.60 per million output tokens; the model is available through the Elements Marketplace add-on.

**Tags:** [AI](https://yomu.fyi/topic/ai), [Heroku](https://yomu.fyi/topic/heroku), [Model Serving](https://yomu.fyi/topic/model-serving), [Open Source](https://yomu.fyi/topic/open-source)

- Source: [Heroku](https://www.heroku.com/blog/heroku-ai-openai-gpt-oss-120b-model-now-available)
- Source URL: https://www.heroku.com/blog/heroku-ai-openai-gpt-oss-120b-model-now-available
- Ingested by Yomu: 2026-09-10T04:04:38.457Z

[Read original post](https://www.heroku.com/blog/heroku-ai-openai-gpt-oss-120b-model-now-available)