New OpenAI Models significantly unveiled, marking the start of collaboration with Groq and HUMAIN for Day Zero release
Groq and HUMAIN Launch OpenAI's Newest Models on GroqCloud
Groq and HUMAIN have announced the immediate availability of OpenAI's two open models, gpt-oss-120B and gpt-oss-20B, on GroqCloud. These models feature full 128K token context windows, real-time inference performance, and integrated server-side tools such as code execution and web search.
Pricing and Availability
The gpt-oss-120B model is priced from approximately $0.15 per million input tokens, while the gpt-oss-20B model is priced from around $0.10 per million input tokens. GroqCloud infrastructure offers high throughput, with the 120B model processing over 500 tokens per second and the 20B model over 1,000 tokens per second. For a limited time, Groq is waiving fees on tool calls made with OpenAI's open models, enhancing cost-efficiency.
The models are hosted on Groq’s optimized inference platform with data centers across North America and sovereign data centers in Saudi Arabia managed by HUMAIN. This ensures global access along with localized support in the Middle East.
Empowering Innovation
This collaboration enables developers worldwide, and especially in Saudi Arabia and the GCC region, to access and deploy these state-of-the-art open models with low latency, real-time responsiveness, and regulatory compliance, supporting both research and enterprise use cases.
Summary
| Model | Pricing (per million input tokens) | Throughput (tokens/second) | Context Window | Availability | Local Support | |----------------|-----------------------------------|----------------------------|----------------|------------------------------|---------------------------| | gpt-oss-120B | ~$0.15 | 500+ | 128K tokens | Globally via GroqCloud | HUMAIN in Saudi Arabia | | gpt-oss-20B | ~$0.10 | 1000+ | 128K tokens | Globally via GroqCloud | HUMAIN in Saudi Arabia |
This offering marks OpenAI's first open-weight language models release in over five years, providing day-zero access and enabling a new wave of AI innovation with cost-effective, high-performance inference.
Tareq Amin, CEO at HUMAIN, stated that Groq delivers unmatched inference speed, scalability, and cost-efficiency for bringing cutting-edge AI to Saudi Arabia. Jonathan Ross, CEO of Groq, stated that Groq was built to run models like these, fast and affordably.
With this launch, OpenAI's newest models are now available worldwide with minimal latency through GroqCloud.
Developers worldwide can now leverage OpenAI's latest models, gpt-oss-120B and gpt-oss-20B, as they're run on GroqCloud, thanks to the integration of artificial-intelligence technology with Groq's infrastructure. This partnership elevates innovation, particularly in Saudi Arabia and the GCC region, by providing premium access to these advanced open models with technology that ensures low latency, real-time responsiveness, and regulatory compliance.
