Google model product
Reasoning-first Gemini preview for agentic coding and complex problem solving
Updated Aug 10, 2026. Default version: Gemini 3.1 Pro Preview
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Gemini 3.1 Pro.
| LiveCodeBench Pro | 2887.0 | 1 | 4 | 100.0% | C | |
| t2-bench | 1.0 | 1 | 23 | 100.0% | C | |
| ARC-AGI v2 | 0.8 | 2 | 16 | 93.3% | C | |
| MMMLU | 0.9 | 2 | 49 | 97.9% | C | |
| SciCode | 0.6 | 2 | 18 | 94.1% | C | |
| APEX-Agents | 0.3 | 3 | 7 | 66.7% | C | |
| GPQA | 0.9 | 3 | 233 | 99.1% | C | |
| LiveBench | 0.8 | 4 | 38 | 91.9% | B | |
| BrowseComp | 0.9 | 9 | 58 | 86.0% | C | |
| SWE-Bench Verified | 0.8 | 9 | 104 | 92.2% | C | |
| FrontierSWE | 0.4 | 10 | 15 | 35.7% | B | |
| Legal Agent Benchmark | 0.0 | 12 | 13 | 8.3% | B | |
| MMMU-Pro | 0.8 | 12 | 65 | 82.8% | C | |
| Terminal-Bench 2.0 | 0.7 | 12 | 49 | 77.1% | C | |
| Finance Agent v2 | 0.4 | 13 | 26 | 52.0% | B | |
| MRCR v2 (8-needle) | 0.3 | 15 | 21 | 30.0% | C | |
| Humanity's Last Exam | 0.5 | 17 | 92 | 82.4% | C | |
| MCP Atlas | 0.7 | 18 | 30 | 41.4% | C | |
| DeepSWE 1.1 | 0.1 | 19 | 19 | 0.0% | B | |
| SWE-Bench Pro | 0.5 | 34 | 44 | 23.3% | C |
Preference and agent-evaluation signals from published Arena datasets.
| agent steerability | overall | 19 | 0.0 | N/A | 88.6K | |
| agent praise complaint | overall | 20 | 0.0 | N/A | 22.5K | |
| agent task outcome explicit | overall | 22 | 0.0 | N/A | 62.1K | |
| agent tool hallucination | overall | 26 | 0.0 | N/A | 2.1M | |
| agent | overall | 28 | -0.0 | N/A | 2.5M | |
| agent bash recovery steps | overall | 38 | -0.1 | N/A | 177.5K |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
The default version has no provider offering with current input or output token prices.
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Gemini 3.1 Pro | 84.3 | N/A | 1M | 65.5K | No | Proprietary | |
| Gemini 3.1 Pro Preview | N/A | N/A | 1M | 65.5K | No | Unspecified |
A concise description based on the published model registry.
Reasoning-first Gemini preview for agentic coding and complex problem solving
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Gemini 3.1 Pro.
Gemini 3.1 Pro's default version was released on Feb 19, 2026.
No official standard PAYG price is currently available for Gemini 3.1 Pro.
Gemini 3.1 Pro is published under Google in the model registry.
The default version has a 1M token context window.
No. The default version is not marked as having publicly available weights.
No published provider offering is currently linked to the default version.
Nearby ranked alternatives include Qwen3.7 Max, Seed 2.1 Pro, GPT-5.6-Terra.