Gemini 3.1 Pro

Leaderboards

Where Gemini 3.1 Pro sits among the top 20 in each category.

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Test results

Every test Gemini 3.1 Pro has a published result on, out of 100%, with its rank among the models that took it.

• GPQA

  #4 / 256

  Graduate-level science questions written to be hard to look up.

  94.3%

  Verified

• SWE-Bench Verified

  #9 / 120

  Fixing real bugs in real GitHub projects, checked by the projects' own tests.

  80.6%

  Self-reported

• Humanity's Last Exam

  #21 / 108

  Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions

  51.4%

  Self-reported · In the text score· In the vision score

• MMMU-Pro

  #13 / 73

  College-level questions that need both an image and text to answer.

  80.5%

  Self-reported · In the vision score

• BrowseComp

  #10 / 68

  Finding hard-to-locate facts by browsing the web.

  85.9%

  Self-reported · In the agents score

• SWE-Bench Pro

  #45 / 61

  Harder, longer software engineering tasks from real codebases.

  54.2%

  Self-reported · In the coding score

• Terminal-Bench 2.0

  #12 / 54

  Getting real work done in a command-line terminal.

  68.5%

  Self-reported

• MMMLU

  #2 / 51

  Multilingual Massive Multitask Language Understanding dataset released by OpenAI, featuring professionally translated MMLU test questions across 14 languages including Arabic, Bengali, German, Spanish, French, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, Swahili, Yoruba, and Chinese. Contains approximately 15,908 multiple-choice questions per language covering 57 subjects.

  92.6%

  Self-reported

• DeepSWE 1.1

  #50 / 50

  DeepSWE 1.1 evaluates software engineering agents using the mini-swe-agent harness.

  12.0%

  Self-reported · In the coding score

• LiveBench

  #4 / 38

  LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.

  79.9%

  Self-reported · In the text score

• MCP Atlas

  #23 / 38

  MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across complex multi-step tasks.

  69.2%

  Self-reported

• MRCR v2 (8-needle)

  #21 / 32

  MRCR v2 (8-needle) is a variant of the Multi-Round Coreference Resolution benchmark that includes 8 needle items to retrieve from long contexts. This tests models' ability to simultaneously track and reason about multiple pieces of information across extended conversations.

  26.3%

  Self-reported · In the text score

• SciCode

  #3 / 30

  SciCode is a research coding benchmark curated by scientists that challenges language models to code solutions for scientific problems. It contains 338 subproblems decomposed from 80 challenging main problems across 16 natural science sub-fields including mathematics, physics, chemistry, biology, and materials science. Problems require knowledge recall, reasoning, and code synthesis skills.

  59.0%

  Self-reported

• Finance Agent v2

  #15 / 28

  Finance Agent v2 is an agentic financial-analysis benchmark from Vals that evaluates models on real-world finance workflows, measuring their ability to retrieve and reason over financial documents, perform multi-step calculations, and produce accurate analyses.

  43.0%

  Self-reported

• HealthBench Hard

  #17 / 24

  A challenging variation of HealthBench that evaluates large language models' performance and safety in healthcare through 5,000 multi-turn conversations with particularly rigorous evaluation criteria validated by 262 physicians from 60 countries

  21.6%

  Verified

• t2-bench

  #1 / 23

  t2-bench is a benchmark for evaluating agentic tool use capabilities, measuring how well models can select, sequence, and utilize tools to solve complex tasks. It tests autonomous planning and execution in multi-step scenarios.

  99.3%

  Self-reported

• ARC-AGI v2

  #3 / 19

  ARC-AGI-2 is an upgraded benchmark for measuring abstract reasoning and problem-solving abilities in AI systems through visual grid transformation tasks. It evaluates fluid intelligence via input-output grid pairs (1x1 to 30x30) using colored cells (0-9), requiring models to identify underlying transformation rules from demonstration examples and apply them to test cases. Designed to be easy for humans but challenging for AI, focusing on core cognitive abilities like spatial reasoning, pattern recognition, and compositional generalization.

  77.1%

  Self-reported · In the text score· In the vision score

• FrontierSWE

  #11 / 16

  FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.

  40.0%

  Self-reported

• Legal Agent Benchmark

  #12 / 15

  The Legal Agent Benchmark (LAB) is Harvey's open-source benchmark for evaluating AI agents on complex, long-horizon legal work. Tasks are scored under an all-pass standard against expert-curated rubrics, where a task passes only if every required rubric criterion (facts, conclusions, citations, structure, and analytical moves) passes.

  0.0%

  Self-reported

• HealthBench (hard)

  #9 / 13

  21.6%

  Unverified

• APEX-Agents

  #6 / 11

  APEX-Agents is a benchmark evaluating AI agents on long horizon professional tasks that require sustained reasoning, planning, and execution across complex multi-step workflows.

  33.5%

  Self-reported

• GPQA (diamond)

  #1 / 10

  Graduate-level Google-Proof Q&A benchmark with paper-aligned full and diamond subsets.

  92.4%

  Unverified

• LiveCodeBench Pro

  #1 / 5

  LiveCodeBench Pro is an advanced evaluation benchmark for large language models for code that uses Elo ratings to rank models based on their performance on coding tasks. It evaluates models on real-world coding problems from programming contests (LeetCode, AtCoder, CodeForces) and provides a relative ranking system where higher Elo scores indicate superior performance.

  96.2%

  Self-reported