# Re-test after every major model update > Model updates can drastically reshuffle market shares, so a win measured last quarter can silently evaporate and the win-rate protocol should re-run within a set window on each major release. - **Affects (funnel stage):** Choice - **Evidence:** reported - **Confidence:** High - **Impact:** 3/3 - **Status:** active - **Last verified:** 2026-09-18 ## Per-platform status Undocumented platforms show "unknown" - honest unknowns, never invented. | Platform | Status | Notes | Source | | --- | --- | --- | --- | | ChatGPT | partial | OpenAI states rankings will evolve, including personalizing merchant rankings further, which supports retesting. No reshuffle magnitude is documented. | [link](https://help.openai.com/en/articles/11128490-shopping-with-chatgpt-search) (verified 2026-08-28) | | Gemini / Google AI Mode | unknown | Not documented by this platform. | - | | Perplexity | unknown | Not documented by this platform. | - | | Copilot | unknown | Not documented by this platform. | - | | Claude | unknown | Not documented by this platform. | - | ## Evidence and sources - [ACES (arXiv:2508.02630)](https://arxiv.org/abs/2508.02630) (2026-07-08) ## How to act on this Track each engine's current model version and re-run the win-rate protocol within a fixed window whenever a major update lands, comparing against the last baseline. ### Related checklist items - [Re-test after every major model update: behavior reshuffles.](https://agentmint.net/tools/agent-selection-checklist/#b-measurement-02) (Selection) ### Related signals - [Defined agent win-rate measurement](https://agentmint.net/signals/agent-win-rate-measurement/) - [Audit position and ordering effects per engine](https://agentmint.net/signals/position-bias-audit/) - [Price sensitivity varies by model](https://agentmint.net/signals/model-specific-price-sensitivity/) ## Notes Based on ACES (arXiv:2508.02630), a controlled agent-shopping simulation, not measured real-world sales.