Scan: Will there be no next Google Gemini Pro model release by September 30, 2026?

date	predicted	market price	confidence
2026-09-01	83%	82%	71%
2026-08-25	68%	72%	58%

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct polymarket_related kalshi_related article_search claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

What is the most recent Gemini Pro model (e.g., gemini-3.1-pro, 3.5-pro) made publicly available before market creation (~Aug 17, 2026), and on what date was it released?
Is any Gemini Pro model currently in Preview status that could plausibly receive a GA promotion (which counts per the rules) before Sep 30, 2026?
What is the historical interval between consecutive Gemini Pro-labeled releases (1.5 Pro, 2.0 Pro, 2.5 Pro, 3 Pro, 3.1 Pro, etc.), and what is the implied base rate of a new Pro release within any given ~6-week window?
Are there credible leaks, LMArena/Vertex AI model garden sightings, or Google statements pointing to a new Gemini Pro model (e.g., gemini-3.5-pro, gemini-4-pro) launching in late August or September 2026?
What Google events (Made by Google, Gemini at Work, Cloud events, Pixel launch) are scheduled between Aug 17 and Sep 30, 2026 where a model launch would be likely?
Have competitors (OpenAI, Anthropic, xAI, Meta) recently released or announced frontier models that would pressure Google to ship a new Pro model quickly?
What do the sibling Polymarket date-bucket markets for this event imply about the distribution of the expected release date?

Planner reasoning

This is a 'no release by date' market created ~Aug 17, 2026 with a ~6-week window to Sep 30. The key drivers are the current state of the Gemini Pro lineup (latest release date, whether a Preview model is awaiting GA promotion), Google's historical Pro release cadence, and any leaks/announcements or scheduled Google events in the window. The Polymarket price and sibling date-bucket markets are the primary anchors, supplemented by news on rumored gemini-3.5-pro/4.0 models.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.4s	1	## This Market's Polymarket Data Will there be no next Google Gemini Pro model release by September 30, 2026? - Current price (probability): 86.50% - 7-day price change: +12.50% - 30-day price change: +42.50% - Total volume: $72,941 (USD notional) - Price range: 44.00% - 90.00% - Data points: 2
polymarket_related	OK	7.2s	0	Scanned 100 active Polymarket markets, kept 0 matches. keyword 'Gemini Pro': 0 markets \| keyword 'Gemini 3.5': 0 markets \| keyword 'Gemini 4': 0 markets \| keyword 'Google Gemini release': 0 markets
kalshi_related	OK	7.1s	1	1 related markets / summaries. keyword 'Gemini': no matches \| keyword 'Google AI model': no matches \| keyword 'Gemini Pro': ok
article_search	OK	5.0s	22	Fetched 22 unique articles across 5 queries (lookback=45d). 'Google Gemini Pro new model release': 10 articles \| 'Gemini 3.5 Pro': 10 articles \| 'Gemini 4 Pro leak': 10 articles \| 'Gemini Pro preview general availability': 10 articles \| 'Google DeepMind Gemini announcement September 2026': 10 articl
claude_news	OK	20.0s	6	Based on my research, here are the key findings: - Current released Gemini Pro model: Gemini 3.1 Pro (released Feb 19, 2026), still in Generative AI Preview status per Google Cloud docs — "This product or feature is a Generative AI Preview offering... Gemini 3.1 Pro is our most advanced re
gdelt_news	FAILED	240.1s	0	timeout after 240.0s
wikipedia	OK	0.4s	2	Fetched 2 Wikipedia entries (0 missing pages).
code_execution	OK	52.6s	0	## Key Findings - Historical cadence data: 8 Gemini Pro-labeled events since Dec 2023 yield "new-family" intervals (1.0→1.5→2.0→2.5→3.0→3.1) of 71, 356, 48, 238, 89 days (mean=160d, median=89d, high variance), and "all-event" intervals (incl. previews/GA) of **71, 89, 267, 48, 84, 154, 89 d

3. Evidence Brief Sonnet 5 · 7503 chars

# Current state Gemini 3.1 Pro (released ~Feb 19, 2026) remains the latest Pro-labeled model publicly available, and it is still officially in "Preview" status per Google Cloud docs. Gemini 3.5 Pro, announced at I/O on May 19, 2026, has missed multiple targeted release windows (June, mid-July, early August) and remained unreleased/"coming soon" as of early September 2026. Google has instead shipped a string of Flash-tier models (3.6, 3.7, 3.8 Flash) but no new Pro-tier or GA-promoted Pro model since 3.1 Pro. # Timeline of key events - 2026-02-19: Gemini 3.1 Pro released (Preview status) — [claude_news/Google Cloud docs, confirmed] - 2026-05-19: Google announces Gemini 3.5 Pro at I/O; no release date given — [claude_news, confirmed] - 2026-06/07 (mid-July): First and second targeted release windows for 3.5 Pro missed — [codersera.com, reported] - 2026-08-04/05: Google announces three new Gemini models (non-Pro); same period, Gemini 3.5 Pro confirmed still delayed per Bloomberg reporting — [CNBC, confirmed] - 2026-08-05: Demis Hassabis steps back to DeepMind chairman; Jeff Dean exits Google; reshuffle explicitly tied by press to stalled model releases/missed deadlines — [CNBC, confirmed] - 2026-08-12: Pixel 11 launch features "most agentic Gemini yet" but no new Pro model; new DeepMind lead Kavukcuoglu says shipping 3.5 Pro is priority #1 — [CNBC, confirmed] - Mid-Aug 2026: Unverified X leak claims Gemini 3.5 Pro internally cancelled — [claude_news, rumored/unconfirmed] - 2026-09-02: Gemini 3.8 Flash released (coding/agentic focus); no Pro model — [CNBC/changelog, confirmed] - Early Sept 2026: Alphabet stock recovering from "longest losing streak in a decade," partly tied to AI model race struggles over summer — [CNBC, confirmed] # Event Will there be no next Google Gemini Pro model release by September 30, 2026? (Resolves YES if no new Pro-labeled model, including GA promotions of existing previews, is publicly available by that date.) # Outcomes to forecast - Yes (no new Pro release by Sep 30, 2026) - No (a new Pro release occurs by Sep 30, 2026) # Kalshi market anchor Polymarket-sourced pricing for this exact market (same ticker cross-listed): YES (no release) at **86.5%**, up sharply from 44% a month ago (+42.5% 30-day, +12.5% 7-day). Volume modest (~$73K total). This is a strong and rising consensus that no Pro release will occur before deadline. No native Kalshi orderbook data was returned; no directly comparable Kalshi markets found (only unrelated FDA markets surfaced under "Gemini Pro" keyword). # Sub-question answers 1. **Most recent Gemini Pro model before market creation** — Gemini 3.1 Pro, released ~Feb 19, 2026, still in "Preview" status per Google Cloud docs [claude_news]. 2. **Any Pro model in Preview that could GA-promote?** — Yes: Gemini 3.1 Pro itself remains in Preview and could technically GA-promote (which would count per rules), but no signal found indicating an imminent GA promotion is planned [claude_news, docs.cloud.google.com]. 3. **Historical interval base rate** — 8 Pro-labeled events since Dec 2023; "new-family" gaps of 71/356/48/238/89 days (median 89d); "all-event" gaps (incl. previews/GA) median 89d also. A 45-day window implies memoryless P(no release) ≈0.68–0.76 depending on definition; conditional models (accounting for 183+ days already elapsed since 3.1 Pro) give ≈0.40–0.73, blended point estimate ~0.55–0.65 [code_execution]. 4. **Leaks/sightings of 3.5 Pro or 4.0 Pro launching Aug–Sep 2026?** — Only a July 2026 benchmark leak (explainx.ai) and an unverified mid-August cancellation rumor on X; no confirmed model ID, pricing, or launch date as of early Sept [claude_news, codersera.com]. 5. **Scheduled Google events Aug 17–Sep 30, 2026** — Pixel 11 launch already occurred Aug 12 (before window) without a new Pro model; no Made by Google/Cloud Next/Gemini at Work event tied to a Pro launch was identified in the window [claude_news]. 6. **Competitive pressure** — Articles reference Google "losing momentum in the AI model race" vs. OpenAI/Anthropic and note Anthropic's cybersecurity lead being targeted by Google's August model releases (non-Pro); no specific new frontier releases from competitors detailed in this window, but general competitive pressure is cited as driving urgency for Google leadership changes [CNBC]. 7. **Sibling Polymarket bucket implications** — No sibling date-bucket markets were found (polymarket_related returned 0 matches); only this single Yes/No market data available. # Key facts (high-confidence, factual) 1. [claude_news/Google Cloud docs] Gemini 3.1 Pro (Feb 19, 2026) is the last Pro release, still Preview-labeled. 2. [CNBC, Aug 5] Gemini 3.5 Pro delayed; Hassabis/Dean exits explicitly linked to model delays. 3. [CNBC, Aug 12] New DeepMind lead Kavukcuoglu's stated first priority is shipping 3.5 Pro. 4. [CNBC/changelog, Sept 2] Only Flash-tier models (3.6/3.7/3.8 Flash) shipped through early September; no Pro release. 5. [Polymarket] YES price rose from 44% to 86.5% over 30 days, reflecting growing market conviction no release will occur. # Cross-market signals - Kalshi related: No directly relevant Kalshi markets found; unrelated FDA markets surfaced only via keyword overlap. - Polymarket: 86.5% YES (no release), strong upward trend, moderate volume ($73K) — treat as primary cross-market anchor. - Sportsbook implied: N/A. # Analyst opinions and speculation - Futurum Group (Nick Patience): New DeepMind lead's top priority should be shipping 3.5 Pro and proving a predictable cadence — implies release not yet imminent as of mid-August. - Constellation Research (Ray Wang): Expects "faster releases, more experimentation" under new leadership — bullish for eventual release but no specific date. - Unverified X leak (mid-Aug): claims 3.5 Pro internally cancelled — unconfirmed, contradicts official "coming soon" messaging. # Directional lean per outcome - **Yes (no release)**: Strong support — repeated delays (3 missed windows), leadership shakeup disrupting roadmap, only Flash-tier ships through Sept 2, cancellation rumor, rising Polymarket consensus (86.5%). Opposing: new leadership publicly prioritizing shipping 3.5 Pro; historical base-rate models suggest a release is "overdue" and could land within window (~35-45% chance per some model specifications); GA promotion of existing 3.1 Pro is a low-bar path that could trigger No resolution. - **No (release occurs)**: Weak support — no confirmed date, model ID, or pricing; no scheduled major event in window; cancellation rumor (even if unconfirmed) adds bearish signal. Slight upside from statistical "overdue" argument and new leadership urgency. # Gaps / unknowns - No confirmation whether Gemini 3.1 Pro will receive a GA promotion (which would resolve as "No"/release) before Sep 30. - No visibility into internal engineering timelines beyond public delay reporting. - Sibling Polymarket date-bucket markets not found — can't triangulate finer distribution. - GDELT news tool failed (timeout), limiting breadth of real-time signal. # Calibration anchors - Kalshi/Polymarket current YES (no release) price: **86.5%**, up from 44% a month ago. - Historical base-rate models (code_execution) suggest P(no release) in this ~45-day window ≈0.55–0.65 (wide range 0.35–0.76 depending on model), notably lower than current market price — suggesting market may be pricing in the specific delay/leadership-chaos narrative more heavily than pure historical cadence would imply.

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

58% conf

Yes	84%
No	16%

Gemini 3.1 Pro (Feb 19, 2026) remains the only Pro model and 3.5 Pro has missed three targeted windows (June, mid-July, early August), with the Aug 5 leadership shakeup and an unverified mid-August cancellation rumor further undermining a near-term ship; through Sept 2 Google shipped only Flash-tier models. The Polymarket/Kalshi anchor of 86.5% YES, rising sharply from 44% a month ago, reflects this narrative and I stay close to it. I shade slightly toward No because the new DeepMind lead has publicly made shipping 3.5 Pro priority #1, historical cadence models put P(no release) closer to 0.55–0.65, and a low-bar GA promotion of the still-Preview 3.1 Pro would resolve the market No. With no confirmed date, model ID, or scheduled launch event in the ~4-week window, YES remains strongly favored.

gpt-5.6-sol

74% conf

Yes	88%
No	12%

The 86.5% YES market anchor is strongly supported by three missed release windows, the leadership shakeup linked to delays, and the absence of a confirmed model ID, launch date, or event before September 30. Gemini 3.1 Pro remains in Preview, so a GA promotion creates a plausible path to No, but there is no evidence that such a promotion is imminent. Historical cadence points to a lower no-release probability, yet it does not capture the model-specific delays, continued Flash-only launches through September 2, and possible cancellation signal. I therefore place YES slightly above the market anchor while retaining meaningful probability for a surprise 3.5 Pro release or 3.1 Pro GA promotion.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasts anchor tightly to the Polymarket 86.5% price without adequately weighting the brief's own observation that historical base-rate models put P(no release) at 0.55–0.65, "notably lower than current market price" — this is a large ~20-30pp gap that both treat as a minor "shade" rather than a substantive conflict requiring more independent reasoning.

Neither forecast meaningfully engages with the GA-promotion path as a distinct, low-bar resolution mechanism — the brief flags that Gemini 3.1 Pro "remains in Preview and could technically GA-promote" with no timeline evidence either way, yet both treat this as a minor caveat rather than modeling it as a quasi-independent probability (e.g., some Google models GA within 6-12 months of preview, and 3.1 Pro is already ~7+ months into Preview by Sept 30).

The unverified X cancellation rumor is used by both as directional evidence for YES, but neither flags the risk that this rumor could be wrong in the other direction — i.e., if 3.5 Pro was merely delayed rather than cancelled, a sudden release (common after leadership shakeups create urgency) could arrive with little public warning, which is a scenario both underweight.

Neither forecaster addresses the discrepancy in market momentum itself: the Polymarket price nearly doubled (44%→86.5%) in 30 days with only "modest" ~$73K volume — this is a thin, fast-moving market that could be prone to overreaction to the leadership-shakeup narrative, yet both forecasters treat the 86.5% figure as a stable, high-confidence consensus rather than interrogating its liquidity/reliability.

Both forecasts converge very closely (0.84 vs 0.88 YES) with GPT-5.6 expressing notably higher confidence (0.74 vs 0.58) despite using nearly identical evidence and reasoning — this convergence-with-divergent-confidence is not explained, and neither justifies why their point estimate should sit above rather than below the anchor given the base-rate tension noted in point 1.

Neither considers the competitive-pressure counter-scenario in enough depth: the brief notes Google is "losing momentum in the AI model race" and leadership is under pressure to ship — this could argue for either a rushed premature release (favoring No) or continued paralysis/caution post-shakeup (favoring Yes), and both forecasts only use this fact one-directionally toward Yes.

6. Aggregator Fable 5.1 · final calibrated forecast

83% (66% confidence)

Yes: 83% No: 17%

The model-specific evidence strongly favors no Pro release: 3.5 Pro has missed three targeted windows, the Aug 5 leadership shakeup was explicitly tied to delays, only Flash-tier models shipped through Sept 2, and there is no confirmed model ID, pricing, date, or launch event in the remaining ~4 weeks. The critique correctly notes the historical cadence base rate (~0.55–0.65 no-release) sits well below the 86.5% market and that the market is thin and fast-moving, so I do not push above the anchor as Forecast 2 did. The GA-promotion path for 3.1 Pro (now 7+ months in Preview, far longer than 2.5 Pro's ~3-month preview) plus the possibility of a rushed 3.5 Pro launch under new leadership pressure justify a slightly larger No allocation than the market implies. Net: YES ~0.83, modestly below the anchor but still strongly favored.

Pipeline Timing

Total pipeline time: 337.0s

Per-tool research timings shown in the Research section above.