Gemini 2.5 Pro

Leaderboards

Where Gemini 2.5 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 2.5 Pro has a published result on, out of 100%, with its rank among the models that took it.

• GPQA

  #83 / 256

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

  83.0%

  Self-reported

• AIME 2025

  #70 / 123

  Competition maths problems from the 2025 American Invitational Mathematics Exam.

  83.0%

  Self-reported

• SWE-Bench Verified

  #87 / 120

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

  63.2%

  Self-reported

• Humanity's Last Exam

  #76 / 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

  17.8%

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

• MMMU

  #17 / 66

  MMMU (Massive Multi-discipline Multimodal Understanding) is a benchmark designed to evaluate multimodal models on college-level subject knowledge and deliberate reasoning. Contains 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and Tech & Engineering across 30 subjects and 183 subfields.

  79.6%

  Self-reported

• AIME 2024

  #5 / 54

  American Invitational Mathematics Examination 2024, consisting of 30 challenging mathematical reasoning problems from AIME I and AIME II competitions. Each problem requires an integer answer between 0-999 and tests advanced mathematical reasoning across algebra, geometry, combinatorics, and number theory. Used as a benchmark for evaluating mathematical reasoning capabilities in large language models at Olympiad-level difficulty.

  92.0%

  Self-reported

• SimpleQA

  #14 / 47

  SimpleQA is a factuality benchmark developed by OpenAI that measures the short-form factual accuracy of large language models. The benchmark contains 4,326 short, fact-seeking questions that are adversarially collected and designed to have single, indisputable answers. Questions cover diverse topics from science and technology to entertainment, and the benchmark also measures model calibration by evaluating whether models know what they know.

  50.8%

  Self-reported

• Aider-Polyglot

  #4 / 22

  A coding benchmark that evaluates LLMs on 225 challenging Exercism programming exercises across C++, Go, Java, JavaScript, Python, and Rust. Models receive two attempts to solve each problem, with test error feedback provided after the first attempt if it fails. The benchmark measures both initial problem-solving ability and capacity to edit code based on error feedback, providing an end-to-end evaluation of code generation and editing capabilities across multiple programming languages.

  76.5%

  Self-reported

• ARC-AGI v2

  #18 / 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.

  4.9%

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

• Video-MME

  #7 / 17

  Video-MME is the first-ever comprehensive evaluation benchmark of Multi-modal Large Language Models (MLLMs) in video analysis. It features 900 videos totaling 254 hours with 2,700 human-annotated question-answer pairs across 6 primary visual domains (Knowledge, Film & Television, Sports Competition, Life Record, Multilingual, and others) and 30 subfields. The benchmark evaluates models across diverse temporal dimensions (11 seconds to 1 hour), integrates multi-modal inputs including video frames, subtitles, and audio, and uses rigorous manual labeling by expert annotators for precise assessment.

  84.8%

  Self-reported

• Global-MMLU-Lite

  #3 / 16

  A lightweight version of Global MMLU benchmark that evaluates language models across multiple languages while addressing cultural and linguistic biases in multilingual evaluation.

  88.6%

  Self-reported

• Aider-Polyglot Edit

  #2 / 10

  A challenging multi-language coding benchmark that evaluates models' code editing abilities across C++, Go, Java, JavaScript, Python, and Rust. Contains 225 of Exercism's most difficult programming problems, selected as problems that were solved by 3 or fewer out of 7 top coding models. The benchmark focuses on code editing tasks and measures both correctness of solutions and proper edit format usage. Designed to re-calibrate evaluation scales so top models score between 5-50%.

  72.7%

  Self-reported

• LiveCodeBench v5

  #1 / 9

  LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from programming contests (LeetCode, AtCoder, CodeForces) and evaluates four different scenarios: code generation, self-repair, code execution, and test output prediction. Problems are annotated with release dates to enable evaluation on unseen problems released after a model's training cutoff.

  75.6%

  Self-reported

• Vibe-Eval

  #2 / 8

  VIBE-Eval is a hard evaluation suite for measuring progress of multimodal language models, consisting of 269 visual understanding prompts with gold-standard responses authored by experts. The benchmark has dual objectives: vibe checking multimodal chat models for day-to-day tasks and rigorously testing frontier models, with the hard set containing >50% questions that all frontier models answer incorrectly.

  65.6%

  Self-reported

• MRCR

  #1 / 7

  MRCR (Multi-Round Coreference Resolution) is a synthetic long-context reasoning task where models must navigate long conversations to reproduce specific model outputs. It tests the ability to distinguish between similar requests and reason about ordering while maintaining attention across extended contexts.

  93.0%

  Self-reported

• MRCR 1M (pointwise)

  #1 / 1

  MRCR 1M (pointwise) is a variant of the Multi-Round Coreference Resolution benchmark that uses pointwise evaluation for ultra-long contexts (~1M tokens). This version evaluates each response independently rather than comparatively, testing models' absolute performance on long-context reasoning tasks.

  82.9%

  Unverified