{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2608-08239","title":"The Replay Gap: Static Evaluation of Model Switching in LLM Agents Scores the Wrong World","arxiv_id":"2608.08239","date":"2026-08-08","proceeding":null,"authors":["Ashritha Gonuguntla"],"abstract":"LLM routers promise efficiency by matching each request to the cheapest adequate model, and are increasingly applied per step inside multi-step agents. Yet agentic routers are evaluated like single-turn routers: by replaying logged trajectories and substituting another model's recorded outputs, assuming the rest of the trajectory is unaffected. We test this assumption with branching rollouts: we fork live SWE-bench agent trajectories at controlled points, rebuild the environment, continue each fork with a different model, and compare against same-model control forks that isolate sampling and replay noise. Across six paired runs (~900 rollouts), swaps exceed their matched control floors by +0.25 to +0.66 normalized edit distance (multiplicity-corrected CIs exclude zero), rewriting 61-94% of post-fork actions; 74-77% of early swaps diverge at the first post-fork action, versus 6-35% of controls, leaving only 3% of replayed states valid. Divergence decreases with fork depth in both directions. All five outcome flips we observe occur in swap arms, upgrades rescuing unsolved instances and a downgrade losing the sole solve, and zero occur across 359 control forks. Scoring these same swaps with a log-stitching replay evaluator, replay mispredicts every success-relevant outcome call and predicts patches with 0.00-0.11 similarity to reality. Auditing the noise floor, temperature-0 \"determinism\" is configuration-dependent: FP8-served controls diverge on over 90% of forks while AWQ-served ones remain near-identical; and under tight budgets the stronger model more often exhausts its steps without submitting. Replay-based benchmarks score the wrong world for agentic routing; we release our harness and all trajectories.","url_abs":"https://arxiv.org/abs/2608.08239","url_pdf":"https://arxiv.org/pdf/2608.08239","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2608.08239","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2608.08239"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/AshrithaG/replay-gap","reach":null}],"summary":{"ran":6},"by_repo_kind":{"found_in_text":{"samples":6,"ran":6,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"3dafafdf0809bf82","entry":"action_divergence","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/metrics.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/metrics.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3dafafdf0809bf82"}},{"code_sha256_prefix":"4eee563f1acfb14a","entry":"assistant_indices","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/branching.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/branching.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4eee563f1acfb14a"}},{"code_sha256_prefix":"2759df64bedf7884","entry":"commands","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/metrics.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/metrics.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2759df64bedf7884"}},{"code_sha256_prefix":"0ec3f8fbc6028884","entry":"levenshtein","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/metrics.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/metrics.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0ec3f8fbc6028884"}},{"code_sha256_prefix":"cf3c08d85122d751","entry":"n_assistant_steps","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/branching.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/branching.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"cf3c08d85122d751"}},{"code_sha256_prefix":"3a5eee7eec29353f","entry":"prefix_for_fork","repo":"AshrithaG/replay-gap","repo_kind":"found_in_text","path":"src/replay_gap/branching.py","file_url":"https://github.com/AshrithaG/replay-gap/blob/HEAD/src/replay_gap/branching.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3a5eee7eec29353f"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}