GoKawiilWeave Router 2.0 matches GPT-6 Astra performance at half the cost
Weave OS released Weave Router 2.0, an open-source model router that dynamically switches between LLMs for coding agents like Claude Code or Codex. On Terminal Bench 4.0 and SWE Atlas benchmarks, the router matched GPT-6 Astra's pass rates while running at roughly 52-54% of its cost and 2.2-2.5x faster. The company attributes the gains to a new routing architecture, a larger training dataset, and improved cache-eviction calculations.
GoKawiil's interpretation of the reporting above, not reported fact.
If the benchmark claims hold up under independent testing, routing between cheaper and pricier models for different subtasks could meaningfully cut the cost of running coding agents at scale, which is a significant expense for companies deploying them widely. The approach also suggests that ensemble routing strategies may rival single frontier models on complex tasks, potentially shifting competitive pressure away from always using the most expensive model available.
- Weave Router 2.0 reportedly matches GPT-6 Astra's pass rates on two coding benchmarks at about half the cost.
- The router dynamically assigns tasks to different models based on complexity, cost, and caching considerations.
- Improvements came from a new architecture, expanded training data, and better cache-eviction impact modeling.
Source: news.ycombinator.com, 2026-09-30
Published there as: “Show HN: Open-source model routing for coding agents at Astra-level performance”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.