Google DeepMindGemini
Gemini 2.5 Pro
Google DeepMind · Gemini. Released Mar 25, 2025.
Gemini 2.5 Pro is a model from Google DeepMind in the Gemini family, released Mar 25, 2025. evals.report tracks 50 reported Gemini 2.5 Pro benchmark scores across LMArena, LiveCodeBench Pro, GPQA Diamond, Humanity's Last Exam, MMMU-Pro, SimpleQA Verified, SWE-bench Verified, FrontierMath, and 42 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
50 results
Benchmark results 50
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| LMArena | Chat preference | 1457 | source-defined rating | Official | Mar 25, 2025 | Details |
| LiveCodeBench Pro | Coding | 1769 | Codeforces Elo | Official | Mar 25, 2025 | Details |
| GPQA Diamond | Reasoning | 85.3% | accuracy | Official | Mar 25, 2025 | Details |
| Humanity's Last Exam | Reasoning | 21.64% | accuracy | Official | Mar 25, 2025 | Details |
| MMMU-Pro | Multimodal | 68.0% | accuracy | Official | Mar 25, 2025 | Details |
| SimpleQA Verified | Other | 56.0% | accuracy | Official | Mar 25, 2025 | Details |
| SWE-bench Verified | Coding | 57.6% | % resolved | Official | Mar 25, 2025 | Details |
| FrontierMath | Reasoning | 14.14% | accuracy | Official | Mar 25, 2025 | Details |
| AIME (OTIS Mock) | Reasoning | 84.2% | accuracy | Official | Mar 25, 2025 | Details |
| WeirdML | Coding | 54.0% | average accuracy | Official | Mar 25, 2025 | Details |
| MCP Atlas | Tool use | 8.8% | pass rate | Official | Mar 25, 2025 | Details |
| Artificial Analysis Intelligence Index | Reasoning | 34.6 | Index | Unverified | Mar 25, 2025 | Details |
| Epoch Capabilities Index | Reasoning | 146.7 | Index | Official | Mar 25, 2025 | Details |
| Aider Polyglot | Coding | 83.1% | % correct | Official | Mar 25, 2025 | Details |
| MMLU-Pro | Reasoning | 86.2% | accuracy | Verified | Mar 25, 2025 | Details |
| GAIA: A Benchmark for General AI Assistants | Agents | 33.3% | accuracy | Unverified | Mar 25, 2025 | Details |
| Video-MME | Multimodal | 84.8% | accuracy | Verified | Mar 25, 2025 | Details |
| GDPval | Agents | 919 | Elo | Official | Mar 25, 2025 | Details |
| LiveCodeBench | Coding | 80.1% | Pass@1 | Unverified | Mar 25, 2025 | Details |
| SciCode | Coding | 42.8% | accuracy | Unverified | Mar 25, 2025 | Details |
| MMMU (Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark) | Multimodal | 81.7% | accuracy | Verified | Mar 25, 2025 | Details |
| IFBench | Reasoning | 52.3% | accuracy | Official | Mar 25, 2025 | Details |
| MultiChallenge | Reasoning | 53.62% | accuracy | Verified | Mar 25, 2025 | Details |
| OpenAI-MRCR v2 (Multi-Round Coreference Resolution) | Reasoning | 58.0% | accuracy (mean SequenceMatcher similarity) | Verified | Mar 25, 2025 | Details |
| LongBench v2 | Reasoning | 63.3% | accuracy | Official | Mar 25, 2025 | Details |
| Global-MMLU | Reasoning | 90.3% | accuracy | Unverified | Mar 25, 2025 | Details |
| Video-MMMU | Multimodal | 83.6% | accuracy | Verified | Mar 25, 2025 | Details |
| WebDev Arena | Chat preference | 1204 | Elo | Verified | Mar 25, 2025 | Details |
| Search Arena | Chat preference | 1143 | Elo | Verified | Mar 25, 2025 | Details |
| Arena-Hard-Auto v2.0 | Chat preference | 79.0% | % win rate | Official | Mar 25, 2025 | Details |
| EQ-Bench Creative Writing v3 | Chat preference | 1417 | Elo | Verified | Mar 25, 2025 | Details |
| Design Arena | Chat preference | 1208 | Elo | Verified | Mar 25, 2025 | Details |
| MASK (Model Alignment between Statements and Knowledge) | Other | 55.67 | Honesty score | Verified | Mar 25, 2025 | Details |
| MCP-Universe | Tool use | 22.08% | Overall Success Rate | Verified | Mar 25, 2025 | Details |
| OCRBench v2 | Multimodal | 59.3 | accuracy | Verified | Mar 25, 2025 | Details |
| ScreenSpot-Pro | Multimodal | 11.4% | accuracy | Verified | Mar 25, 2025 | Details |
| FACTS Grounding | Reasoning | 74.2% | Grounding accuracy | Verified | Mar 25, 2025 | Details |
| BigCodeBench | Coding | 36.5% | calibrated Pass@1 | Verified | Mar 25, 2025 | Details |
| EnigmaEval | Reasoning | 5.57% | accuracy | Verified | Mar 25, 2025 | Details |
| ZeroBench | Multimodal | 3.0% (pass@1) | accuracy | Verified | Mar 25, 2025 | Details |
| IMO-Bench | Reasoning | 68.2% | accuracy | Verified | Mar 25, 2025 | Details |
| PutnamBench | Reasoning | 3 | Problems solved | Verified | Mar 25, 2025 | Details |
| FrontierMath Tier 4 | Reasoning | 4.2% | accuracy | Official | Mar 25, 2025 | Details |
| Vectara Hallucination Leaderboard | Other | 7.0% | Hallucination Rate | Official | Mar 25, 2025 | Details |
| Gray Swan Arena (Agent Red-Teaming / Indirect Prompt Injection) | Agents | 8.5% | Attack Success Rate (ASR) | Verified | Mar 25, 2025 | Details |
| PolyMath: Evaluating Mathematical Reasoning in Multilingual Contexts | Reasoning | 52.2 | Difficulty-Weighted Accuracy (DW-ACC) | Verified | Mar 25, 2025 | Details |
| Vibe Code Bench | Coding | 0.40% | Overall accuracy | Verified | Mar 25, 2025 | Details |
| WebArena | Agents | 54.8% | Task success rate | Verified | Mar 25, 2025 | Details |
| GSO: Software Optimization Benchmark for SWE-Agents | Coding | 3.92% | Opt@1 | Official | Mar 25, 2025 | Details |
| MultiNRC | Reasoning | 45.12% | accuracy | Official | Mar 25, 2025 | Details |