Gemini 3.1 Pro preview is built for tasks where simple answers are not enough. Stronger core reasoning for complex coding, math, and long-context workflows, with multimodal support and a reported 77.1% verified score on ARC-AGI-2. NOTE: Inputs > 200k tokens are charged at 2x input and 1.5x output rates.
Added Feb 19, 2026
Context Window
1M
Max Output
65.5K
Avg output tokens (7d)
975 tokens
Input Price (Auto)
$2.00/1M
Output Price (Auto)
$12.00/1M
Cache Read (Auto)
$0.20/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
29.7
Coding Index
68.8
Agentic Index
8.2
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
35.4%
Better than 54% of models compared
Harvey LAB-AA
Legal agentic work criterion pass rate
58.9%
Better than 8% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
453 Elo
Better than 21% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
776 Elo
Better than 39% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
17.8%
Better than 67% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
82.0%
Better than 93% of models compared
MLCR-AA
Medical long-context reasoning
15.6%
Better than 52% of models compared
Reasoning
HLE
Humanity's Last Exam
47.0%
Better than 95% of models compared
IFBench
Instruction-following benchmark
77.1%
Better than 96% of models compared
CritPt
Research-level physics reasoning
17.7%
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
4.0%
Better than 55% of models compared
SciCode
Python programming for scientific computing
58.7%
Better than 92% of models compared
Knowledge
AA-Omniscience Accuracy
Proportion of correctly answered questions
54.8%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
50.9%
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
94.1%
Better than 99% of models compared
Terminal-Bench Hard (legacy)
Agentic coding and terminal use
53.8%
Better than 96% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
95.6%
Better than 95% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
82.0%
Better than 93% of models compared
GDPval-AA (unversioned / legacy)
Economically valuable tasks
13.8%
Last updated Sep 29, 2026
Artificial AnalysisProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…
Related text models
Compare Gemini 3.1 Pro (Preview) with similar models from the same provider or model family.
Gemini Pro Latest
google/gemini-pro-latestCompatibility alias that routes to the newest version of Gemini Pro. Currently routes to Gemini 3.1 Pro Preview High.
Gemini 3.1 Pro (Preview Custom Tools)
google/gemini-3.1-pro-preview-customtoolsGemini 3.1 Pro preview variant tuned for better tool selection behavior in coding agents and multi-tool workflows. It reduces overuse of generic bash tools and improves function-calling reliability while retaining Gemini 3.1 Pro's multimodal reasoning and 1M-token context. NOTE: Inputs > 200k tokens are charged at 2x input and 1.5x output rates.
Gemini 3.1 Pro (Preview High)
google/gemini-3.1-pro-preview-highGemini 3.1 Pro preview high-reasoning variant.
Gemini 3.1 Pro (Preview Low)
google/gemini-3.1-pro-preview-lowGemini 3.1 Pro preview low-reasoning variant.
Gemini 3.8 Flash
google/gemini-3.8-flashGoogle's fast multimodal model for agentic workloads, including coding, tool use, image understanding, PDF and document extraction, audio, and video. Its capabilities, limits, reasoning behavior, and pricing currently mirror Gemini 3.7 Flash.
Gemini 3.7 Flash
google/gemini-3.7-flashGoogle's frontier-performance Flash model for multimodal and agentic workloads, including coding, tool use, image understanding, PDF and document extraction, audio, and video. Google reports 65.3% on DeepSWE v1.1, up from 49.0% for Gemini 3.6 Flash, and 34% on GDP.pdf, up from 14%.