Gemini 3.1 Pro
PreviewModel Gemini 3.1 Pro is a preview model by Google: the provider has listed it as a preview since 19 Feb 2026. A general release is not confirmed.
- Where can it be used?
- Evidenced as usable in: GeminiRemoved from: GitHub CopilotFor 12 more recorded products there is no statement; that does not mean “not available”.Available with limits in the provider’s API
- Did we test it?
- No lab run published yet
- Last change
- 7 months ago
- Evidence
- 5 sources from 4 sitesRegistry entry:
checked 11 Oct 2026
What backs this up?
Three kinds of evidence, kept apart. None of them proves AGI on its own; where one is missing, the page says so.
- In real productsEvidenced as usable in: Gemini · Removed from: GitHub Copilot
- agi.jetzt PrüfstandNo published run yet.
- Public benchmarks20 reported results on 17 benchmarks — each with its source
A model and the product around it are different things: a result for one does not carry over to the other.
Public benchmark results
Results reported by benchmark maintainers or providers, each with its source. They are not averaged and not compared across benchmarks; a provider result is self-reported. Some sources are data files of the benchmark’s own site rather than pages; they are marked as machine-readable.
- Gemini 3.1 Pro26.1 %Chartography · v1 · accuracy (expert-range graded), no toolssource as of 03/10/2026 · benchmark maintainer · mode not stated · judge: google/gemini-3.5-flash · tools: no · setting: Inspect AI (surge-ai/chartography) · attempts: 10 · web: no · reasoning: High · computer use: no · reported as „Gemini 3.1 Pro (High reasoning)“Source: surgehq.ai — Chartography, Gemini 3.1 Pro
- Gemini 3.1 Pro17 %GDP.pdf · v1 · pass rate, all criteria (strict)source as of 03/10/2026 · benchmark maintainer · mode not stated · judge: google/gemini-3.5-flash · tools: no · setting: Inspect AI (surge-ai/gdp-pdf) · attempts: 5 · web: no · reasoning: High · computer use: no · reported as „Gemini 3.1 Pro (High reasoning)“Source: surgehq.ai — GDP.pdf, Gemini 3.1 Pro
- Gemini 3.1 Pro$911.21Vending-Bench 2 · 2 · final bank balancesource as of 03/10/2026 · benchmark maintainer · model + harness · Andon Labs Vending-Bench 2 agent environment (email, web search, inventory tools, notes/reminders) · tools: yes · budget: ~1 simulated year; 3000-6000 messages, 60-100M output tokens per run · attempts: 5 · web: yesMachine-readable source: andonlabs.com · data bundle of the website, not linked
- Gemini 3.1 Pro77.08 %ARC-AGI-2 · semi-private · accuracy (task pass rate)source as of 02/10/2026 · benchmark maintainer · mode not stated · reasoning: thinking (preview) · reported as „Gemini 3.1 Pro (Preview)“Machine-readable source: arcprize.org · JSON — ARC-AGI-2, Gemini 3.1 Pro
- Gemini 3.1 Pro3.62 %CL-Bench (Continual Learning Bench) · 2026-07-18 · normalized gain (vs own baseline) · iclsource as of 18/07/2026 · benchmark maintainer · model alone · setting: in-context learning only (CL-Bench "ICL" system) · reported as „ICL · Gemini 3.1 Pro Preview“Machine-readable source: continual-learning-bench.com · JSON — CL-Bench (Continual Learning Bench), Gemini 3.1 Pro
- Gemini 3.1 Pro8.1 %CL-Bench (Continual Learning Bench) · 2026-07-18 · normalized gain (vs own baseline) · icl notepadsource as of 18/07/2026 · benchmark maintainer · model + harness · ICL Notepad (memory scaffold, CL-Bench) · reported as „ICL Notepad · Gemini 3.1 Pro Preview“Machine-readable source: continual-learning-bench.com · JSON — CL-Bench (Continual Learning Bench), Gemini 3.1 Pro
- Gemini 3.1 Pro26.83 %FrontierMath Tier 4 · v2 · accuracy (mean score)evaluated 11/06/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 1,000,000 tokens hard limit (forced submission at 660,000) · setting: Epoch AI FrontierMath implementation (Inspect): python tool + submit_answer tool · web: no · reported as „gemini-3.1-pro-preview“Machine-readable source: epoch.ai · ZIP download — FrontierMath Tier 4, Gemini 3.1 Pro
- Gemini 3.1 Pro59.65 %FrontierMath Tiers 1–3 · v2 · accuracy (mean score)evaluated 11/06/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 1,000,000 tokens hard limit (forced submission at 660,000) · setting: Epoch AI FrontierMath implementation (Inspect): python tool + submit_answer tool · web: no · reported as „gemini-3.1-pro-preview“Machine-readable source: epoch.ai · ZIP download — FrontierMath Tiers 1–3, Gemini 3.1 Pro
- Gemini 3.1 Pro6.4 hMETR time horizon (50 % success) · 1.1 · time horizonsource as of 08/05/2026 · benchmark maintainer · model + harness · triframe_inspect (Inspect) · tools: yes · reported as „gemini_3_1_pro“Machine-readable source: metr.org · YAML — METR time horizon (50 % success), Gemini 3.1 Pro
- Gemini 3.1 Pro1.5 hMETR time horizon (80 % success) · 1.1 · time horizonsource as of 08/05/2026 · benchmark maintainer · model + harness · triframe_inspect/triframe_agent · tools: yes · reported as „gemini_3_1_pro (Gemini 3.1 Pro)“Machine-readable source: metr.org · YAML — METR time horizon (80 % success), Gemini 3.1 Pro
- Gemini 3.1 Pro26.03 %τ³-bench banking (pass@1) · tau2-bench-1.0.1 · pass@1evaluated 05/05/2026 · benchmark maintainer · model + harness · τ²-bench harness (Sierra, v1.0.1) · tools: yes · attempts: 4 · pass@k: 1 · reasoning: high · reported as „Gemini 3.1 Pro Preview“Machine-readable source: raw.githubusercontent.com · JSON — τ³-bench banking (pass@1), Gemini 3.1 Pro
- Gemini 3.1 Pro9.28 %τ³-bench banking (pass^4 consistency) · tau2-bench-1.0.1 · pass^4 (all 4 runs succeed)evaluated 05/05/2026 · benchmark maintainer · model + harness · τ²-bench harness (Sierra, v1.0.1) · tools: yes · attempts: 4 · pass@k: pass^4 · reasoning: high · reported as „Gemini 3.1 Pro Preview“Machine-readable source: raw.githubusercontent.com · JSON — τ³-bench banking (pass^4 consistency), Gemini 3.1 Pro
Show 8 more
- Gemini 3.1 Pro46.44 %Humanity's Last Exam · 2025-04 · accuracyevaluated 10/04/2026 · benchmark maintainer · mode not stated · judge: o3-mini-2025-01-31 · reasoning: thinking high · temperature: 0.0 when configurable · reported as „gemini-3.1-pro-preview (thinking high)“Source: labs.scale.com — Humanity's Last Exam, Gemini 3.1 Pro
- Gemini 3.1 Pro32.2 %SWE-bench Pro · private · resolve rateevaluated 08/04/2026 · benchmark maintainer · model + harness · mini-swe-agent · tools: yes · reasoning: thinking · reported as „gemini-3.1-pro (thinking)*“Source: labs.scale.com — SWE-bench Pro, Gemini 3.1 Pro
- Gemini 3.1 Pro46.1 %SWE-bench Pro · public · resolve rateevaluated 08/04/2026 · benchmark maintainer · model + harness · mini-swe-agent · tools: yes · reasoning: thinking · reported as „gemini-3.1-pro (thinking)*“Source: labs.scale.com — SWE-bench Pro, Gemini 3.1 Pro
- Gemini 3.1 Pro84.9 %MRCR v2 (8-needle), Google DeepMind · 8-needle · score at 128k contextsource as of 19/02/2026 · self-reported by the provider · mode not stated · context: 128k · reported as „Gemini 3.1 Pro Thinking (High)“Source: deepmind.google — MRCR v2 (8-needle), Google DeepMind, Gemini 3.1 Pro
- Gemini 3.1 Pro26.3 %MRCR v2 (8-needle), Google DeepMind · 8-needle · score at 1M contextsource as of 19/02/2026 · self-reported by the provider · mode not stated · context: 1M · reported as „Gemini 3.1 Pro Thinking (High)“Source: deepmind.google — MRCR v2 (8-needle), Google DeepMind, Gemini 3.1 Pro
- Gemini 3.1 Pro80.5 %MMMU-Pro · v1 · accuracy, no toolssource as of 19/02/2026 · self-reported by the provider · model alone · reasoning: Thinking (High) · reported as „Gemini 3.1 Pro Thinking (High)“Source: deepmind.google — MMMU-Pro, Gemini 3.1 Pro
- Gemini 3.1 Pro79.6 %SimpleBench · avg5 · accuracy (average of 5 runs)evaluated 17/02/2026 · benchmark maintainer · model alone · pass@k: AVG@5 · temperature: 0.7 (top-p 0.95) · reported as „Gemini 3.1 Pro Preview“Machine-readable source: simple-bench.com · JavaScript file — SimpleBench, Gemini 3.1 Pro
- Gemini 3.1 Pro0.42 %ARC-AGI-3 · semi-private · accuracyhistoricalbenchmark maintainer · mode not stated · reported as „Gemini 3.1 Pro (Preview)“Machine-readable source: arcprize.org · JSON — ARC-AGI-3, Gemini 3.1 Pro
Where can it be used?
Only recorded availability statements are shown. A missing plan or surface means unknown, not unavailable. Where it applies only to a plan or a region, the row says so. “Rolling out” is not “available”. “Rollout reported” means a rollout was reported, but nothing has supported it for more than 90 days.
| Product | Status | since | Checked | Sources |
|---|---|---|---|---|
| GeminiApp · AI Pro | Limited availability | not documented | checked 11 Oct 2026 | 1gemini.google |
| GeminiApp · AI Ultra | Limited availability | not documented | checked 11 Oct 2026 | 1gemini.google |
| GeminiApp · Free | Limited availability | not documented | checked 4 Oct 2026 | 1gemini.google |
| GitHub Copilot | Removed | checked 3 Oct 2026 | 1docs.github.com |
Where a start is “not documented”, a source check found the state in place; the source names no start date.
No statement on record for 12 further products: ChatGPT, ChatGPT Work, Claude, Claude Code, Claude Enterprise, Claude Team, Codex, Gemini for Google Workspace, Grok, Langdock, Le Chat Enterprise, Microsoft 365 Copilot. No statement is not the same as not available.
Developer APIs
Developer access — not a statement about any work product.
| Product | Status | since | Checked | Sources |
|---|---|---|---|---|
| Gemini APIAPI | Limited availability | not documented | checked 3 Oct 2026 | 1ai.google.dev |
What did we test?
No lab run of Gemini 3.1 Pro is published yet.
Nothing is shown until a real run exists. See what the lab tests:
Lifecycle and history
The dates the provider confirmed, and every recorded period of availability. A current period shows its checked status; one without a source check is marked not verified.
Announced
Preview since
Deprecated—not recorded
Retired—not recorded
What changed?
Sourced changes concerning this model, newest first.
- 7 months agoadded to the record on 29 Sept 2026
- 7 months agoadded to the record on 29 Sept 2026
Evidence
Every source document behind the facts on this page, by domain.
| Domain | Docs | Documents |
|---|---|---|
| gemini.google | 2 | |
| ai.google.dev | 1 | |
| blog.google | 1 | |
| docs.github.com | 1 |