
- Who
- Microsoft Research, UC Berkeley, and Microsoft Security
- What
- introduced and evaluated TokenEconomizer, an open-source plugin that uses repository session history to reduce coding agent costs
- Quote
TokenEconomizer lowered the average cost by 18.4–30.9%, as shown in Figure 3, and saved the most in all 30 repository-combination pairs without performance loss.
What happened
Microsoft Research and collaborators introduce TokenEconomizer, an open-source plugin evaluated on 277 SWE-Bench tasks across six repositories and five AI coding agent and model combinations. It reduced average costs by 18.4–30.9%, with cumulative costs falling below the default in every repository by the sixth task. The paper also reports lower costs on sequential web development and security auditing.
Why it matters
Coding teams can reduce agent costs across repeated work in the same repository, with savings appearing by the sixth task even when the system starts without session history.
Read next
See the full storyMicrosoft Research and Collaborators Release TokenEconomizer to Cut AI Coding Agent CostsThe story so far
Microsoft Research, UC Berkeley, and Microsoft Security introduced TokenEconomizer, an open-source plugin that uses repository session history to cut AI coding agent costs, and reported average cost reductions of 18.4–30.9% without performance loss in an evaluation of 277 SWE-Bench tasks. Across six repositories and five agent-and-model combinations, the plugin saved the most in all 30 repository-combination pairs, with cumulative costs falling below the default in every repository by the sixth task after starting with no prior session history. The researchers said TokenEconomizer also lowered costs on sequential web development and security auditing. Designed to run alongside existing AI coding agents without modifying their underlying models, the plugin requires no API key of its own and keeps what it learns on the developer's machine.
