Microsoft and others release TokenEconomizer to reduce coding agent costs
Understand this story
UC Berkeley, Microsoft Research, and the Microsoft Security team have released an open-source plugin called TokenEconomizer to reduce the token costs of AI coding agents. It turns historical sessions that the agent has already completed in a repository into guidance for subsequent sessions, reducing repeated exploration, loading useless tools, and getting stuck in failure loops. The plugin does not modify the agent or its underlying model, nor does it require its own API key. The learned content is stored locally on the developer's machine, so it can be used with existing coding agents.
AI overview based on source reports · Updated 1 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Microsoft ResearchFeaturedTokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient
TokenEconomizer is an open-source plugin from UC Berkeley, Microsoft Research, and Microsoft Security that learns from an agent's past sessions in a repository to cut waste without modifying the agent or its model. Across 277 SWE-Bench tasks and five agent-model combinations, it reduced average cost by 18.4–30.9% in all 30 repository-combination pairs, with task success rates broadly stable. It works through micro-skills, tool pruning, and failure-loop intervention, and also lowered costs 16.0–21.2% on Web-Bench and 6.4–19.6% on security auditing.
Activity for this story
Not enough continuous observations to show a trend yet.