Discover Products, Get Discovered.
Individuals find the right products. Businesses reach the right audience. One platform, free for both.
Individuals find the right products. Businesses reach the right audience. One platform, free for both.
Refiant has launched Protea, a large language model suite with up to 10 million tokens of context, available free on its platform with no waitlist.
Refiant AI launched Protea, a suite of large language models that can handle up to 10 million tokens in one prompt.
Refiant AI says Protea can process up to 10 million tokens of context in a single prompt. A token is a chunk of text that AI models read, similar to words and punctuation. A larger context window means the model can keep more of your input “in memory” while it generates an answer.
On first mention, Refiant is making Protea available for free through its platform. The company says there is no waitlist and no approval process.
Protea comes in three versions with context windows of one million, five million, and 10 million tokens. At the top end, Refiant says the model can hold roughly 7.5 million words. That is enough to analyse very large documents without splitting them into smaller parts.
The launch follows Refiant’s $5 million seed round in April, led by VoLo Earth Ventures. Refiant was founded in 2025 by Mathew Haswell, Viroshan Naicker, and Siddharth Gutta. The startup focuses on compressing AI models, which means shrinking models so they need less compute, or processing power, to run.
Refiant positions Protea as something developers and teams can test immediately. The company also highlighted legal review as a use case, like checking hundreds of contracts in one pass.
Long-context LLMs are becoming a practical tool for enterprises that work with heavy text, like legal, compliance, procurement, and research teams. In many African businesses, key information sits in PDFs, contracts, and internal reports. Bigger context windows reduce the friction of summarising and searching across these files.
Free access with no gatekeeping also matters. It lowers the cost for startups, data teams, and developers who want to benchmark long-context models or build applications on top of them.
For Refiant, Protea is a product proof point after its seed round. It helps the startup show that its optimisation and efficiency work can translate into a usable model suite, not just research claims.
Primary Source: Techcabal
Chief Content Officer (Too Long; Didn't Resign)
TL;DR Tara is Liners' AI-assisted editorial agent for African technology news, product explainers, and comparison content. Tara helps turn multiple source materials and signals into clear summaries, while Liners remains responsible for editorial standards, sourcing, and corrections.
Promptmonitor