Nous Research
@NousResearch
Qwen 3.7 Max is now supported in Hermes Agent https://t.co/cM5SqoJxu3
50retweets805likes
View on XNous Research Adds Qwen 3.7 Max Support to Hermes Agent
Nous Research· Updated
Nous Research integrated Alibaba's Qwen 3.7 Max into its open-source Hermes Agent platform. This allows users to power autonomous multi-step workflows with reasoning models while benefiting from recent cost-saving context caching.
Nous Research, an organization focused on open-source AI, added official support for Alibaba's Qwen 3.7 Max to its Hermes Agent platform. Hermes Agent is an autonomous system that plans and executes multi-step tasks using an agent loop (an iterative cycle of reasoning and action).
This integration provides an open-source alternative to proprietary tools. It arrives alongside Qwen's context caching feature, which reduces the cost of the massive system prompts required for agentic workflows by 90 percent. By combining frontier reasoning with these economics, Hermes Agent becomes more viable for long-horizon engineering tasks.
You can access the model through the Nous Portal gateway or by connecting a DashScope API key. The update also supports browser-based OAuth sign-in, similar to the Grok subscription portability feature recently added to the platform. Configuration is handled via the hermes model command or the config.yaml file.
Still wondering? A few quick answers below.
Hermes Agent is an open-source autonomous AI system developed by Nous Research that can plan and execute multi-step tasks. Unlike standard chatbots, it uses an agent loop to reason about problems, use external tools, and navigate complex environments like codebases to complete goals without constant human direction.
You can connect Qwen 3.7 Max to Hermes Agent through several official paths. The recommended method is the Nous Portal subscription, but you can also use a DashScope API key for direct cloud access or sign in via browser-based OAuth using a consumer Qwen Portal account for individual use.
Yes, Hermes Agent is designed to work with any OpenAI-compatible API endpoint, including local providers. You can run open-weight models on your own hardware using tools like Ollama, vLLM, or llama.cpp and point the agent to your local server address to maintain data privacy and avoid per-token costs.
Hermes Agent requires a Unix-based environment (like Linux or macOS) and at least 64,000 tokens of context length to function reliably as an agent. This large window is necessary because the system must fit the complex system prompt, tool definitions, and conversation history into memory to maintain reasoning across multi-step workflows.
The platform includes a fallback provider configuration that automatically switches to a backup model if the primary one fails due to rate limits or server errors. This chain of backup providers ensures that autonomous tasks can continue without interruption by swapping the reasoning engine mid-session while preserving the conversation state.
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →
