Research

Paper

AI LLM March 03, 2026

Evaluating Performance Drift from Model Switching in Multi-Turn LLM Systems

Authors

Raad Khraishi, Iman Zafar, Katie Myles, Greig A Cowan

Abstract

Deployed multi-turn LLM systems routinely switch models mid-interaction due to upgrades, cross-provider routing, and fallbacks. Such handoffs create a context mismatch: the model generating later turns must condition on a dialogue prefix authored by a different model, potentially inducing silent performance drift. We introduce a switch-matrix benchmark that measures this effect by running a prefix model for early turns and a suffix model for the final turn, and comparing against the no-switch baseline using paired episode-level bootstrap confidence intervals. Across CoQA conversational QA and Multi-IF benchmarks, even a single-turn handoff yields prevalent and statistically significant, directional effects and may swing outcomes by -8 to +13 percentage points in Multi-IF strict success rate and +/- 4 absolute F1 on CoQA, comparable to the no-switch gap between common model tiers (e.g., GPT-5-nano vs GPT-5-mini). We further find systematic compatibility patterns: some suffix models degrade under nearly any non-self dialogue history, while others improve under nearly any foreign prefix. To enable compressed handoff risk monitoring, we decompose switch-induced drift into per-model prefix influence and suffix susceptibility terms, accounting for ~70% of variance across benchmarks. These results position handoff robustness as an operational reliability dimension that single-model benchmarks miss, motivating explicit monitoring and handoff-aware mitigation in multi-turn systems.

Metadata

arXiv ID: 2603.03111
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-03
Fetched: 2026-03-04 03:41

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.03111v1</id>\n    <title>Evaluating Performance Drift from Model Switching in Multi-Turn LLM Systems</title>\n    <updated>2026-03-03T15:44:57Z</updated>\n    <link href='https://arxiv.org/abs/2603.03111v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.03111v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Deployed multi-turn LLM systems routinely switch models mid-interaction due to upgrades, cross-provider routing, and fallbacks. Such handoffs create a context mismatch: the model generating later turns must condition on a dialogue prefix authored by a different model, potentially inducing silent performance drift. We introduce a switch-matrix benchmark that measures this effect by running a prefix model for early turns and a suffix model for the final turn, and comparing against the no-switch baseline using paired episode-level bootstrap confidence intervals. Across CoQA conversational QA and Multi-IF benchmarks, even a single-turn handoff yields prevalent and statistically significant, directional effects and may swing outcomes by -8 to +13 percentage points in Multi-IF strict success rate and +/- 4 absolute F1 on CoQA, comparable to the no-switch gap between common model tiers (e.g., GPT-5-nano vs GPT-5-mini). We further find systematic compatibility patterns: some suffix models degrade under nearly any non-self dialogue history, while others improve under nearly any foreign prefix. To enable compressed handoff risk monitoring, we decompose switch-induced drift into per-model prefix influence and suffix susceptibility terms, accounting for ~70% of variance across benchmarks. These results position handoff robustness as an operational reliability dimension that single-model benchmarks miss, motivating explicit monitoring and handoff-aware mitigation in multi-turn systems.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-03T15:44:57Z</published>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>Raad Khraishi</name>\n    </author>\n    <author>\n      <name>Iman Zafar</name>\n    </author>\n    <author>\n      <name>Katie Myles</name>\n    </author>\n    <author>\n      <name>Greig A Cowan</name>\n    </author>\n  </entry>"
}