$NBIS announced a partnership with LangChain to integrate Nebius Token Factory with LangChain’s Deep Agents. The goal is to make it easier f…

$NBIS announced a partnership with LangChain to integrate Nebius Token Factory with LangChain’s Deep Agents. The goal is to make it easier for teams building AI agents on LangChain to run those workloads on Nebius infrastructure using open-source models. The integration combines: • LangChain Deep Agents for orchestration • LangSmith for tracing and evaluation • Nebius Token Factory as the inference backend • Tavily for real-time web search and content extraction This matters because agentic workflows usually require more infrastructure than simple chatbot interactions. A single agent workflow can involve planning, sub-agent calls, tool use, memory retrieval, retries, and multiple inference requests. Nebius Token Factory supports 30+ open-source models, including Llama, Qwen, DeepSeek, and NVIDIA Nemotron. It also offers an OpenAI-compatible API, dedicated endpoints, autoscaling, and a 99.9% uptime SLA. One useful part of the integration is that teams can route different sub-agents to different models. For example, they can use a larger model for planning, a faster model for execution, and a lighter model for tool-calling. This can help optimize cost and performance across the agent workflow. The integration also supports embeddings and semantic retrieval through the langchain-nebius package, meaning teams can run more of the agent pipeline on Nebius infrastructure. Overall, this is another step in Nebius positioning Token Factory as infrastructure for production AI workloads, especially for teams that want to build agents using open models.

Source: Harrison Chase (X) | 2026-05-12