forecasting
agent
← back to scans

Will the next Google Gemini Pro model added to the Arena Leaderboard debut at a score of at least 1490?

polymarketCompanies·2026-10-12·view on polymarket ↗
28%
Agent
28%
Market Price
-0.4%
Edge
48%
Confidence
60,300
Volume
0.8c
Spread
80
Days to resolution
5
Markets in event
final rationale
Mean of 2 draws (0.27, 0.28). (0) Not resolved. The Arena changelog through Oct 9 and the Oct 8 leaderboard show no newly added Pro-labeled Gemini since gemini-3.1-pro-preview in February. The only Gemini 4 entry, gemini-4-argon-high (added Sept 30, score 1525), has no 'Pro' label, so under the market's strict labeling rule it almost certainly does not qualify. (a) About 80 days remain until the Dec 31, 2026 deadline. (b) If nothing changes, no Pro-labeled Gemini is added and the market resolves No by default. (c) NO: 3.5 Pro stays shelved, and Google ships further Gemini 4 variants (Carbon, a GA Argon) under non-Pro names, or nothing new reaches Arena before year-end. (d) YES: Google releases a 'Gemini 4 Pro' or a late 3.5 Pro onto Arena in Nov/Dec. Given Argon's 1525 and a frontier band of about 1505–1531, it would very likely debut at 1490 or higher. Evidence for that path includes Kavukcuoglu's 'much earlier than year-end' comment, introductory API pricing that runs through Dec 31, and Business Insider's report of Carbon testing. The sibling strikes from 1480 to 1510 all trade near 27–31%, so the market sees appearance, not score, as the binding constraint. I agree with that view. The 54-point drop over 30 days reflects Argon taking the flagship slot without the Pro name. I see no unpriced information strong enough to move far from the 27.9% anchor.
price vs. prediction
market price agent forecast at t+0
6 follow-ups
re-scan context
This market has been scanned before. Previous predictions:
datepredictedmarket priceconfidence
2026-10-04 37% 36% 52%
2026-09-24 87% 92% 61%
1. Planner Opus 5.5 · picks the research tools
Tools chosen: polymarket_direct resolution_source polymarket_related kalshi_related article_search
Planner reasoning
polymarket_direct is the price anchor. resolution_source on the Arena text leaderboard shows current top scores, including where existing Gemini Pro models sit and whether a new Pro model has already appeared. polymarket_related and kalshi_related add cross-market signal on Gemini release timing and top-model scores, and earnings_data on GOOGL is a light structured check.
2. Research parallel tool calls + 4 web research loops (Sonnet 5.5) · read raw by the forecasters
at a glance
RESEARCH DIGEST (built in code: tool summaries and each report's bottom line; the forecasters read the full tool outputs and reports) Tools - polymarket_direct: ## This Market's Polymarket Data **Will the next Google Gemini Pro model added to the Arena Leaderboard debut at a score of at least 1490?** - Current price (probability): 27.90% (live; bid 28% / ask 28%) - 7-day price change: -2.95% - 30-day price change: -53.85% - Total volume: $60,300 (USD notional) - Price range: 27.65% - 94.75% - Data points: 90 days - resolution_source: already_resolved=unclear; matches_criteria=True; last_updated=Oct 8, 2026 (leaderboard); Oct 9, 2026 (changelog). The Text Arena Overall leaderboard (dated Oct 8, 2026; 414 models) does not show a newly added Gemini model labeled "Pro". The newest Gemini text model is gemini-4-argon-high, which ranks #1 at 1525 ±9 (Preliminary), but its name has no "Pro" label. According to the changelog, the last - polymarket_related: Polymarket search over 4 keyword(s), kept 18 matches. keyword 'Gemini': 5 markets | keyword 'Arena leaderboard': 5 markets | keyword 'top AI model': 3 markets | keyword 'Google AI model': 5 markets - kalshi_related: 1 related markets / summaries. keyword 'Gemini': no matches | keyword 'top AI model': ok | keyword 'LMArena': no matches - article_search: Fetched 12 unique articles across 2 queries (lookback=45d). 'Gemini 3.5 Pro release date': 8 articles | 'Gemini Pro LMArena score debut': 8 articles Research loops, bottom lines - current state: The key unknown is whether gemini-4-argon-high (1525, added Sept 30, not labeled "Pro") counts. It likely doesn't, and Gemini 3.5 Pro or a Gemini 4 Pro-labeled model is still pending with no date. If a Pro model arrives, its score would likely be near or above 1490 given the current top-of-board scores. The main risk to Yes is that no qualifying model is added before Dec 31, 2026. - base rate: All three prior Pro-labeled Gemini debuts scored 1470–1501, and they have scored higher with each generation. Frontier scores are now about 1505–1525, so a new Pro debuting at 1490 or above looks likely if one appears. The main risk is that no Pro-labeled Gemini is added by Dec 31, because 3.5 Pro is stalled, Argon is not labeled Pro, and any later Pro-labeled Gemini 4 model is unconfirmed. - yes case: The strongest YES path is a Gemini 4-family model (Carbon, or an Argon-tier release with a "Pro" label) or a late 3.5 Pro appearing on Arena by Dec 31. Such a model would very likely score at least 1490, given Argon's 1525 and the 1500–1531 top band. The weak point is timing and the "Pro" label, not the score. - no case: The most likely path to No is that no Pro-labeled Gemini is added by Dec 31. Argon already took the flagship slot under a non-Pro name, 3.5 Pro is stalled or dropped, and Google has not confirmed any Gemini 4 Pro name. A strict label reading of any later Gemini 4 entry adds a second route to No.
tools
toolstatustimeitemssummary
polymarket_direct OK 0.5s 1 ## This Market's Polymarket Data **Will the next Google Gemini Pro model added to the Arena Leaderboard debut at a score of at least 1490?** - Current price (probability): 27.90% (live; bid 28% / ask 28%) - 7-day price change: -2.95% - 30-day price change: -53.85% - Total volume: $60,300 (USD notio
resolution_source OK 16.7s 1 already_resolved=unclear; matches_criteria=True; last_updated=Oct 8, 2026 (leaderboard); Oct 9, 2026 (changelog). The Text Arena Overall leaderboard (dated Oct 8, 2026; 414 models) does not show a newly added Gemini model labeled "Pro". The newest Gemini text model is gemini-4-argon-high, which rank
polymarket_related OK 0.8s 18 Polymarket search over 4 keyword(s), kept 18 matches. keyword 'Gemini': 5 markets | keyword 'Arena leaderboard': 5 markets | keyword 'top AI model': 3 markets | keyword 'Google AI model': 5 markets
kalshi_related OK 2.3s 1 1 related markets / summaries. keyword 'Gemini': no matches | keyword 'top AI model': ok | keyword 'LMArena': no matches
article_search OK 0.6s 12 Fetched 12 unique articles across 2 queries (lookback=45d). 'Gemini 3.5 Pro release date': 8 articles | 'Gemini Pro LMArena score debut': 8 articles
research loops · 4/4 ok · 14 searches · 3 page reads
dimensionstatustimesearchespage readsreport
current state OK 20s 4 0 4073 chars
base rate OK 22s 4 0 4846 chars
yes case OK 21s 4 3 4280 chars
no case OK 19s 2 0 5009 chars
report: current state
- [2026-09-30] Arena's leaderboard changelog shows gemini-4-argon-high was added to the Text, Agent and WebDev leaderboards on this date. Its name has no "Pro" label. It is Google's most recent frontier addition, so it probably does not count as the "next Gemini Pro model" under the rules. How the market's resolver treats it is unclear. https://arena.ai/company/leaderboard-changelog - [2026-09-30] Arena posted that Gemini 4 Argon (High) debuted #1 in Text Arena with 1525 points. That is +20 over #2 Claude Opus 4.6 (High) and well above Google's previous release, Gemini 3.8 Flash (High) at #11. https://x.com/arena/status/2105394855644139908 - [2026-10-08] The live Text leaderboard (overall, style control off) still lists gemini-4-argon-high at #1 with 1525 ±9 and marks it "Preliminary". It has 4,892 votes and costs $2/$10 per million tokens. Claude Opus 5.5 high is #2 at 1507. The top of the board is about 1500–1525. https://arena.ai/leaderboard/text - [2026-09-30] Google announced Gemini 4 Argon, limited to vetted cyber defenders in the Fairwind program. It does not appear in the public Gemini API list as of Oct 6, and Google has given no date for wider release. https://www.rdworldonline.com/googles-overdue-gemini-4-argon-reaches-the-frontier-with-mixed-results-for-science/ - [2026-10-03] Gemini 3.5 Pro never shipped. It was announced at I/O on May 19 with a "next month" target, and it missed June, mid-July and early August. The "3.5 Pro coming soon" label was still on Google's Pro page on Oct 3, with no model card or pricing. Google has not formally cancelled it. https://aiwiz.uk/blog/gemini-3-5-pro/ - [2026-07-21] Google said Gemini 3.5 Pro was in partner testing and would be broadly available when ready. A Bloomberg report, cited secondhand, said internal performance goals had delayed it. https://windowsforum.com/windows-news.4/gemini-3-5-pro-no-launch-date-or-verified-lm-arena-test.441354/ - [2026-07-31] Community posts said gemini-3.5-pro briefly appeared in Arena blind testing and was removed. This is unverified, since the sources are X and Discord posts and one outlet. A Gemini 3.5 Pro leaderboard entry has not been confirmed. https://x.com/AiBattle_/status/2083071686039961860 - [2026-02-19] The last Pro model on record is Gemini 3.1 Pro, which Wikipedia lists as the current stable Pro. A third-party tracker put Gemini 3.1 Pro Preview at about 1505 in August. Gemini 3 Pro debuted at 1501 in Nov 2025, per Yahoo/Google. https://en.wikipedia.org/wiki/Google_Gemini - [2026-10-06] Gemini 3.8 Flash (released Sept 2) was replaced in Arena's battle pool by Argon. Google's latest Flash-tier models sit well below the top: Gemini 3.6 Flash was #23 and 3.8 Flash was #11 on Arena text. This suggests Flash models don't qualify and Pro-tier scores are what matter. https://x.com/ash_twtz/status/2107432082041602359 - [2026-10-09] Recent changelog entries (Oct 2–9) contain no Gemini Pro addition. The only Google additions were Argon (Sept 30) and nano-banana-2.1 (image, Oct 6). No Pro-labeled Gemini entry is visible as of Oct 9. https://arena.ai/company/leaderboard-changelog - [2026-10-12] No announced or scheduled Gemini Pro release date was found before the market closes on 2027-01-01. The scores above show that a Pro model on par with Argon would clear 1490, and that the recent top-of-board Gemini Pro scores are around 1500. https://arena.ai/leaderboard/text BOTTOM LINE: The key unknown is whether gemini-4-argon-high (1525, added Sept 30, not labeled "Pro") counts. It likely doesn't, and Gemini 3.5 Pro or a Gemini 4 Pro-labeled model is still pending with no date. If a Pro model arrives, its score would likely be near or above 1490 given the current top-of-board scores. The main risk to Yes is that no qualifying model is added before Dec 31, 2026. NOT FOUND: any announced release date for a Gemini 3.5 Pro or Gemini 4 Pro model, or any Arena "Pro"-labeled Gemini entry added after Feb 2026. I also found no confirmation that Argon is formally a "Pro" tier, and no official Arena statement on 3.5 Pro testing.
report: base rate
- [2026-10-12] Gemini 4 Argon is not labeled "Pro." Arena's changelog shows `gemini-4-argon-high` was added on 2026-09-30 to the Text, Agent and WebDev boards. The market covers only models labeled "Pro", so Argon probably doesn't count. I could not confirm how the live board labels it. Another page says Argon landed Oct 6, which conflicts with the changelog's Sep 30 entry. https://arena.ai/company/leaderboard-changelog - [2026-09-30] Arena posted that Gemini 4 Argon (High) landed #1 in Text with 1525 points, about 20 above Claude Opus 4.6 (High). Its previous Google release, Gemini 3.8 Flash (High), was #11. If a Gemini 4 "Pro" ships later, this is the nearest comparable score. https://x.com/arena/status/2105394855644139908 - [2026-10-03] Gemini 3.5 Pro was promised for June 2026 and never shipped. Google announced Gemini 4 Argon on Sep 30. The 3.5 Pro "coming soon" label was still up on Oct 3, and 3.5 Pro was not in the API model list. Pichai said on the Jul 22 earnings call that Google "need[s] Gemini 4 as a larger base model". In September Kavukcuoglu said Google had stepped back from 3.5 Pro to focus on Flash. SemiAnalysis reported in August that 3.5 Pro had been quietly dropped; that is a report, not confirmed. https://aiwiz.uk/blog/gemini-3-5-pro/ - [2026-10-12] The newest Pro on the board is Gemini 3.1 Pro Preview. Its release was announced Feb 19, 2026. Google's own Gemini page lists 3.1 Pro as the stable Pro release. No new Pro model has appeared on the Arena changelog through Oct 9. https://en.wikipedia.org/wiki/Google_Gemini - [2026-02-19] Base rate, Gemini 3.1 Pro: it debuted at 1500 in Text and was tied #1. Arena described it as 4 points behind Opus 4.6. Third-party tables show a 1489 figure, probably an earlier or style-control reading. https://x.com/arena/status/2024519891295089063 - [2025-11-18] Base rate, Gemini 3 Pro: it debuted at 1501 and was #1 in Text, about 50 points above Gemini 2.5 Pro. https://x.com/arena/status/1990813759938703570 - [2025-03-25] Base rate, Gemini 2.5 Pro: debuted #1, with a record +40 point jump over the prior leader. Its score was about 1440–1470; a third-party recap lists 1470, which I could not verify. https://x.com/lmarena_ai/status/1904581128746656099 - [2025-12-10] Gemini 3 Pro's score drifted to 1492 after about 15.9k votes, and one tracker shows 1486 by spring 2026. Scores can fall 10–15 points after debut. The market reads the score at 12:00 PM ET the day after first appearance, so this drift mostly doesn't apply. https://felloai.com/the-best-ai-of-december-2025/ - [2026-10-12] Score inflation is limited. Opus 4.6 was ~1503 in spring. Opus 5.5 High was 1507 (±8) on Oct 8, and Argon is 1525. Gemini 3.8 Flash (High) debuted at 1494, and Flash 3.7 was 1491. Even Flash models in the Gemini 3.x line were at or above ~1490. A Google Pro model released after Argon would likely be at least on par with the frontier. https://x.com/arena/status/2095179977805517149 - [2026-10-08] Per The Sweat, the Oct 8 board shows Argon at 1525 ±9, ahead of Claude Opus 5.5 High at 1507 ±8. Argon's score rests on only about 4.9k votes. https://thesweat.com/predictions/best-ai-model-of-2026/ - [2026-10-12] Timing is the main uncertainty. Resolution needs a new Pro-labeled Gemini to appear by Dec 31, 2026. Past Pro-labeled gaps were about 7 months (2.5 → 3) and about 3 months (3 → 3.1). 3.5 Pro has been dropped or delayed, and Google's flagship effort is now Gemini 4. The "Pro" tier name for Gemini 4 is unconfirmed. I found no official statement of a Gemini 4 Pro or a date for one. - [2026-10-12] Third-party aggregator chanceindex lists sibling strikes for this market, but the page is undated and may be stale. It shows ≥1480 at 91.7%, ≥1495 at 70.0%, ≥1500 at 39.5% and ≥1510 at 23.6%, with 1490 at 75.9%. It also says the headline contract ranged 30–83% over its life. It reflects the pre-Argon environment. https://chanceindex.com/odds/next-google-gemini-pro-model-arena-debut - [2026-10-12] The Manifold market (Oct 8) for the #1 company on Oct 31 had Google at 86%. It does not address the Pro-model question. https://manifold.markets/Terminator2/which-company-will-have-the-1-model BOTTOM LINE: All three prior Pro-labeled Gemini debuts scored 1470–1501, and they have scored higher with each generation. Frontier scores are now about 1505–1525, so a new Pro debuting at 1490 or above looks likely if one appears. The main risk is that no Pro-labeled Gemini is added by Dec 31, because 3.5 Pro is stalled, Argon is not labeled Pro, and any later Pro-labeled Gemini 4 model is unconfirmed. NOT FOUND: Any official announcement of a Gemini 4 Pro or other new Pro model with a date. The Arena label for Argon on the live board. Rumors of a Pro model debuting on Arena under a codename. Polymarket's own historical resolution record for this contract.
report: yes case
YES means: the next Gemini model labeled "Pro" to be newly added to the Arena Text leaderboard (style control off) shows a Score of 1490 or higher at 12:00 PM ET on the day after it first appears, and it appears by Dec 31, 2026. - [2026-10-09] Business Insider reported that Google employees are testing a newer Gemini 4 build, codenamed Carbon, internally. One employee said it "feels like Opus 5.5" for coding. The report says it is unclear whether Carbon would ship under the Argon name or as a separate Gemini 4 model. This is the most concrete sign of a further Gemini 4 model, and a follow-on model could be labeled "Pro". Source: https://cellcog.ai/blog/gemini-4-release-date/ - [2026-09-24] Kavukcuoglu said Gemini 4 was in early post-training and would launch "much earlier" than year-end, with an early post-training version out "as soon as possible". This is reported from The Information and 9to5Google, with no Google transcript. It supports a Gemini 4 wider release before Dec 31. Source: https://cellcog.ai/blog/gemini-4-release-date/ - [2026-09-30] Google said wider availability goes first to paid API customers and AI Ultra subscribers, with no date. Argon API pricing is introductory through Dec 31, 2026 and then doubles. That footnote implies Google plans a broad release inside 2026. The Arena listing for a "Pro" model could follow that release. Source: https://meetcody.ai/blog/gemini-4-argon-gemini-pro-delay/ - [2026-10-01] The leaderboard changelog entry reads "gemini-4-argon-high". The earlier Arena entry for Gemini 3.8 Flash reads "Gemini 3.8 Flash (High)". Neither carries a "Pro" label. Argon is the only Google text-model addition since Sept 2. If the resolver counts the Gemini 4-tier flagship as the "Pro" tier, it was already added on Sept 30 at 1525. That would be a Yes. This route is weak, because the question text requires a "Pro" label and one source says there is "no 'Pro' this time". Sources: https://news.lmarena.ai/leaderboard-changelog/ and https://getneotiler.com/blog/gemini-4-pro-release-date-leaked-specs-and-whats-actually-confirmed/ - [2026-09-17 to 09-26] Community posts and several outlets called the stealth "gemini-3.8-flash" Arena checkpoint a "Gemini 4 Pro" test. This is rumor, but the same checkpoint series then debuted at 1525 as Argon. That shows the Gemini 4 line tests well in Arena blind voting, with a 30-point margin over the 1490 threshold. Source: https://cellcog.ai/blog/gemini-4-release-date/ - [2026-08] The top of the board is in a tight 1500–1531 band. A third-party tracker put Claude Mythos 5 at 1531 on Oct 6. Any Pro that is competitive at the frontier lands above 1490. Even Gemini 3.8 Flash debuted at 1494 (per earlier reports). Source: https://www.swfte.com/lmarena - [2026-03] Gemini 3.1 Pro landed 3rd in Text Arena, just behind Claude Opus 4.6. This confirms that Pro-labeled Gemini debuts have sat at or near the top. Source: https://arena.ai/blog/march-2026-arena-updates - [2026-07-21] Google's own post said Gemini 3.5 Pro was "testing with partners" and would be broadly available when ready. It has not been cancelled, and a Jul 31 Arena sighting is unverified. A short-notice 3.5 Pro debut would need to score about 1490, a bar a 3.1 Pro successor should clear. Sources: https://cellcog.ai/blog/gemini-4-release-date/ and https://x.com/AiBattle_/status/2083071686039961860 - [2026-10-06] Arena's policy adds a preliminary score at launch for models released in Arena Direct Chat. Restricted-access models such as Argon (Fairwind only) were therefore listed without a public API. The same route lets a Pro debut on Arena before general availability, before Dec 31. Source: https://arena.ai/blog/policy BOTTOM LINE: The strongest YES path is a Gemini 4-family model (Carbon, or an Argon-tier release with a "Pro" label) or a late 3.5 Pro appearing on Arena by Dec 31. Such a model would very likely score at least 1490, given Argon's 1525 and the 1500–1531 top band. The weak point is timing and the "Pro" label, not the score. NOT FOUND: any Google statement that a Gemini 4 "Pro" tier exists, any Arena entry labeled Pro since Feb 2026, any dated Carbon or 3.5 Pro release, and the Business Insider page itself (secondhand via cellcog.ai). I could not fetch pasqualepillitteri.it.
report: no case
YES means: the next Gemini model labeled "Pro" to be newly added to the Arena text leaderboard shows an overall score of 1490 or higher at 12:00 PM ET on the day after it first appears. This requires such a model to appear by Dec 31, 2026. - [2026-10-12] Arena's changelog, fetched in full, lists no Pro-labeled Gemini from Sept 5 to Oct 9. Gemini additions in that window are only gemini-4-argon-high (Sept 30) and nano-banana-2.1 (an image model, Oct 6). Arena also adds models in batches that mix Text, Agent, WebDev and Vision entries. A Pro-labeled Gemini would show up there as a separate entry. https://arena.ai/company/leaderboard-changelog - [2026-10-08] Argon is the only confirmed Gemini 4 variant. As of Oct 7, no official source confirms a Gemini 4 Pro, Flash or other variant. That means the next Gemini 4 entry on Arena may carry a non-"Pro" name, and the Polymarket text excludes anything "labeled only as" a non-Pro variant. Runtimewire (Oct 8) says Google has given no date for wider access and has not confirmed whether Pro app tiers get it. https://www.consens.io/topics/gemini-4-release-date and https://runtimewire.com/article/everything-we-know-about-google-s-gemini-4-argon - [2026-10-12] CellCog's tracker (Oct 11) says Gemini 4 "Pro" and "Flash" are leaker and community names with no Google product name. It also says Kavukcuoglu told The Information on Sept 23 that launch would be "much earlier" than year-end. The release-date estimates it cites are Nov/Dec from 9to5Google and TestingCatalog, and a rumored September launch did not happen. Argon is already on Arena, so a later general-availability release of Argon would probably not be a new entry. https://cellcog.ai/blog/gemini-4-release-date/ - [2026-10-03] Google's summer pattern points away from Pro. It skipped 3.5 Pro and shipped Flash-tier models on July 21 and Sept 2, and Kavukcuoglu said in September that it had stepped back from 3.5 Pro. Runtimewire says Google "skipped 3.5 Pro over the summer", and aiwiz says a 3.5 Pro launch "now looks unlikely". The "coming soon" label has no model card or pricing. https://aiwiz.uk/blog/gemini-3-5-pro/ and https://runtimewire.com/article/everything-we-know-about-google-s-gemini-4-argon - [2026-08-02] The July 31 report of "gemini-3.5-pro" on Arena rests on X and Discord posts and one outlet. WindowsForum noted that no preserved result, screenshot or statement from Google or Arena backs it. A model withdrawn from blind testing never reached the leaderboard, so that sighting does not count as an addition. https://windowsforum.com/windows-news.4/gemini-3-5-pro-no-launch-date-or-verified-lm-arena-test.441354/ - [2026-10-12] Arena publishes Google's top-end models under idiosyncratic names, such as "gemini-4-argon-high" with no "Pro". A future top-tier Gemini 4 model could do the same, even if Google markets it as the Pro tier. The market's resolver might read "labeled as Pro" strictly. Argon's own announcement had no tier at all, and Google says Argon is the "frontier model". The market text names gemini-2.5-pro, gemini-3-pro and gemini-3.1-pro-preview as examples. https://arena.ai/company/leaderboard-changelog - [2026-10-12] Deadline arithmetic: Argon had limited access from Sept 30 and no API date (Oct 5 trackers), with a U.S. government voluntary pre-release review still in progress per Runtimewire (Oct 8). Arena listed Argon at once, so Arena addition is not tied to public availability. Even so, the roughly 80 days left would have to cover a separate Pro-labeled model. Google released no Pro model in the 7.5 months after 3.1 Pro (Feb 19). https://runtimewire.com/article/everything-we-know-about-google-s-gemini-4-argon - [2026-10-12] Polymarket's own page (about 25 days old when indexed) describes 3.5 Pro as in "internal testing and limited preview" and sets the "no qualifying release by Dec 31" clause to No. It shows that a No by default is built into the contract's design. https://polymarket.com/event/next-google-gemini-pro-model-arena-debut - [2026-10-12] Score risk, secondary. If a Pro did arrive, it could be a smaller or cheaper tier below Argon. Google's recent non-flagship models debuted around 1491–1494 (Flash 3.7 and 3.8, per colleagues), which is only 1–4 points above the strike. The 1490 strike has little cushion there, and Arena's "Preliminary" scores can move. https://arena.ai/leaderboard/text BOTTOM LINE: The most likely path to No is that no Pro-labeled Gemini is added by Dec 31. Argon already took the flagship slot under a non-Pro name, 3.5 Pro is stalled or dropped, and Google has not confirmed any Gemini 4 Pro name. A strict label reading of any later Gemini 4 entry adds a second route to No. NOT FOUND: Any Google statement or leak of a dated Pro-labeled Gemini release or Arena codename in October–December 2026. Any official Arena or Google confirmation of the 3.5 Pro sighting. Whether Arena's live board labels Argon as "Pro". I could not retrieve the full list of changelog entries before Sept 5.
3. Forecasters 2 draws · Opus 5.5 + GPT · plain mean
claude-opus-5-5 · high effort
50% conf
Yes 27%
No 73%
(0) Not resolved. The Arena changelog through Oct 9 and the Oct 8 leaderboard show no newly added Pro-labeled Gemini since gemini-3.1-pro-preview in February. The only Gemini 4 entry, gemini-4-argon-high (added Sept 30, score 1525), has no 'Pro' label, so under the market's strict labeling rule it almost certainly does not qualify. (a) About 80 days remain until the Dec 31, 2026 deadline. (b) If nothing changes, no Pro-labeled Gemini is added and the market resolves No by default. (c) NO: 3.5 Pro stays shelved, and Google ships further Gemini 4 variants (Carbon, a GA Argon) under non-Pro names, or nothing new reaches Arena before year-end. (d) YES: Google releases a 'Gemini 4 Pro' or a late 3.5 Pro onto Arena in Nov/Dec. Given Argon's 1525 and a frontier band of about 1505–1531, it would very likely debut at 1490 or higher. Evidence for that path includes Kavukcuoglu's 'much earlier than year-end' comment, introductory API pricing that runs through Dec 31, and Business Insider's report of Carbon testing. The sibling strikes from 1480 to 1510 all trade near 27–31%, so the market sees appearance, not score, as the binding constraint. I agree with that view. The 54-point drop over 30 days reflects Argon taking the flagship slot without the Pro name. I see no unpriced information strong enough to move far from the 27.9% anchor.
claude-opus-5-5 · high effort
45% conf
Yes 28%
No 72%
(0) The condition has not been met. The resolution-source read (leaderboard Oct 8, changelog Oct 9) shows no Pro-labeled Gemini added since gemini-3.1-pro-preview in Feb 2026. gemini-4-argon-high (1525, added Sep 30) carries no "Pro" label. The market's collapse from about 80% to 28% over 30 days shows traders concluding that Argon does not count. (a) About 80 days remain until the Dec 31, 2026 deadline. (b) If nothing changes, no qualifying model appears and the contract resolves No by default. (c) No: Google keeps the Argon/codename naming for Gemini 4 and leaves 3.5 Pro shelved. Any later Gemini 4 entry (Carbon, or a GA version of Argon) appears under a non-Pro name or not at all before year end. Google has not shipped a Pro in the 7.5 months since 3.1 Pro. (d) Yes: Google ships a broad Gemini 4 release branded "Gemini 4 Pro", or a late 3.5 Pro, and Arena adds it by December. Given Argon at 1525 and frontier scores of 1505–1531, such a model would very likely debut at 1490 or above. Hints in that direction are the introductory pricing through Dec 31, Kavukcuoglu's "much earlier than year-end" remark, and the reported Carbon build. The sibling strikes at 1480, 1500 and 1510 all trade around 28–31%. That confirms the market treats the score as near-certain if a Pro appears, and prices the probability of a Pro-labeled release at about 30%. I found no unpriced information, so I stay at the anchor.
4. Final plain mean of 2 draws
28% (48% confidence)
Yes: 28% No: 72%
Mean of 2 draws (0.27, 0.28). (0) Not resolved. The Arena changelog through Oct 9 and the Oct 8 leaderboard show no newly added Pro-labeled Gemini since gemini-3.1-pro-preview in February. The only Gemini 4 entry, gemini-4-argon-high (added Sept 30, score 1525), has no 'Pro' label, so under the market's strict labeling rule it almost certainly does not qualify. (a) About 80 days remain until the Dec 31, 2026 deadline. (b) If nothing changes, no Pro-labeled Gemini is added and the market resolves No by default. (c) NO: 3.5 Pro stays shelved, and Google ships further Gemini 4 variants (Carbon, a GA Argon) under non-Pro names, or nothing new reaches Arena before year-end. (d) YES: Google releases a 'Gemini 4 Pro' or a late 3.5 Pro onto Arena in Nov/Dec. Given Argon's 1525 and a frontier band of about 1505–1531, it would very likely debut at 1490 or higher. Evidence for that path includes Kavukcuoglu's 'much earlier than year-end' comment, introductory API pricing that runs through Dec 31, and Business Insider's report of Carbon testing. The sibling strikes from 1480 to 1510 all trade near 27–31%, so the market sees appearance, not score, as the binding constraint. I agree with that view. The 54-point drop over 30 days reflects Argon taking the flagship slot without the Pro name. I see no unpriced information strong enough to move far from the 27.9% anchor.
Pipeline Timing
Total pipeline time: 55.2s
Per-tool research timings shown in the Research section above.