Weave Released Router 2.0 to Optimize AI Inference
The new tool routes coding prompts between models to cut inference costs by approximately half.
Updated on Sept. 23, 2026 in Artificial Intelligence
Live Poll
Is now a good time for your organization to adopt automated AI model routing tools?
Weave has launched the Weave Router 2.0, a source-available software tool that evaluates prompt complexity to distribute requests across various AI models. The platform is designed to improve cost and performance for coding agents.
Why it matters
The tool addresses the growing need for efficient infrastructure in AI-driven development by automatically selecting models based on task requirements. It aims to optimize expensive inference workflows that have become a standard operational burden for startups and developers.
The router utilizes a classifier to gauge prompt complexity, paired with cache-aware switching to identify potential cost savings before selecting a model. It supports providers such as Claude, Codex, and GPT while offering potential inference spend cuts of approximately 50%.
The players
Weave
A software company focused on infrastructure tools for coding agents and AI-driven development workflows.
The details
The software functions as an intelligent gateway that intercepts incoming coding requests and assesses their difficulty level. By employing a classifier — an algorithm trained to categorize data points — it routes simpler tasks to smaller, lower-cost models while directing complex prompts to more capable, high-performance engines. A cache-aware switching mechanism further ensures that requests are optimized for both speed and expense by checking existing cached responses.
Timeline
September 23, 2026: Weave Router 2.0 was officially launched.
The Tech Race
The release of Weave Router 2.0 follows a broader industry trend toward source-available distribution under the Elastic License 2.0 for developer-facing tools. It represents a pivot toward granular cost management in an ecosystem currently dominated by monolithic model access.
Solo developers and startups can integrate the router to manage their inference costs, with pricing set at 5% of routed inference spend. Teams of five or fewer can utilize a free starter plan, while larger organizations can opt for the Pro Plan at $50 per engineer per month.
The takeaway
The Weave Router 2.0 offers a tactical method for developers to lower their operational AI expenses through automated intelligent routing. Watch for future benchmarks regarding how the classifier handles edge-case prompts compared to manual model selection.
Further reading
For more on the evolving software layer behind generative agents, explore our Artificial Intelligence section.
Source note: This article includes information reported by Dynamic Business.
Live Poll
Is now a good time for your organization to adopt automated AI model routing tools?
