According to a recent LinkedIn post from Together AI, the company is making NVIDIA’s Nemotron 3 Super model available on its AI Native Cloud platform. The post highlights that this open hybrid mixture-of-experts model is designed for multi-agent systems and complex reasoning, with 120 billion parameters but only 12 billion active per inference step to balance performance and efficiency.
Claim 55% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The LinkedIn post describes a hybrid Mamba-transformer architecture that is suggested to deliver more than 50% faster throughput than leading open models, alongside a 1 million-token context window for extended reasoning across documents and tasks. It also notes benchmark strength in coding, cybersecurity, finance, search, and retail, and emphasizes production-oriented features such as a 99.9% SLA and dedicated infrastructure.
From an investor perspective, the integration of Nemotron 3 Super into Together AI’s cloud offering may enhance the platform’s value proposition for enterprise and developer customers seeking advanced reasoning and agentic AI capabilities. This could support higher usage, stickier workloads, and potentially improved pricing power as customers adopt more compute-intensive models.
The post also references collaboration with NVIDIA AI, which may signal a deepening relationship with a major ecosystem partner and access to cutting-edge GPU-optimized models. If this collaboration expands, Together AI could strengthen its competitive position in AI infrastructure, differentiate against other model-hosting providers, and capture incremental demand from sectors such as finance and cybersecurity where high-performance, long-context models are increasingly important.