New Model, Old Habits: Behavior Transfer under Model Swapping in Multi-Session LLM Agents
Factual evidence
What the source reports
Researchers show that swapping LLMs in multi-session agents causes recipient models to adopt reasoning patterns from past session histories.
Inspect the evidence
- Inclusion basis
- Enterprise AI
- Publisher and source type
- arXiv cs.LG — Machine Learning · RESEARCH
- Published by source
- 15 October 2026
- Collected by OneBench
- 16 Oct 2026, 03:02 UK
Stored source excerpt
arXiv:2610.19712v1 Announce Type: new Abstract: Multi-session agents can switch between LLMs to balance performance and inference cost. Such switches require transferring a history for continuity,…
Short excerpt from the collected text, not the full source. Use the source link to read it in context.
The factual summary is a OneBench synthesis, not a quotation or independent verification. Collection time is not publication time. Open the source for its full context; related reporting can share the same underlying announcement.
OneBench interpretation
Institutional assessment
So what
Model swapping in agentic architectures introduces behavioral drift and unexpected reasoning patterns across sessions, complicating model risk governance.
Do what
Review agent orchestration designs with the model risk team to evaluate behavioral risks when swapping underlying LLMs.