Rankings / Google DeepMind
Gemini 3.1 Pro Preview
released 2026-02-19
BenchAtlas Index
as of 2026-07-14
base configuration
rank #18
12 families · 5 categories · high
high effort
rank #77
5 families · 4 categories · medium
high effort
rank #228
7 families · 4 categories · medium
Benchmark evidence
74 results
agentic coding
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort | 45.4 % | independent | LiveBench |
| τ²-Bench | 95.6 % | independent | Artificial Analysis | |
| τ²-Bench subset=banking | 16.5 % | independent | Artificial Analysis |
coding
| Release 2026-06-25 tasks_counted=2 · livebench_version=2026-06-25 | high effort | 76.5 % | independent | LiveBench |
| SciCode | 58.9 % | independent | Artificial Analysis |
data analysis
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort | 78.5 % | independent | LiveBench |
external indices
| AA Coding Index | 68.8 points | independent | Artificial Analysis | |
| Intelligence Index v4.1 | 46.5 points | independent | Artificial Analysis | |
| ECI | 154.8 points | independent | Epoch AI Benchmarking Hub | |
| Vals Index | high effort | 53.8 points | independent | Vals AI |
factuality
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | 42.4 % 41.1–43.7 | independent | Scale Labs | |
| SimpleQA Verified | 77.3 % | independent | Epoch AI Benchmarking Hub 2026-02-19 |
human preference
| Chinese (style control)older version arena=text · category=chinese · style_control=true | 1527.5 1516.9–1538.1 | community | LMArena Leaderboard Dataset | |
| Coding (style control)older version arena=text · category=coding · style_control=true | 1521.0 1515.4–1526.6 | community | LMArena Leaderboard Dataset | |
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1512.5 1507.6–1517.5 | community | LMArena Leaderboard Dataset | |
| Expert (style control)older version arena=text · category=expert · style_control=true | 1511.6 1503.5–1519.8 | community | LMArena Leaderboard Dataset | |
| Industry Life And Physical And Social Science (style control)older version arena=text · category=industry_life_and_physical_and_social_science · style_control=true | 1507.4 1501.0–1513.7 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1506.2 1501.8–1510.6 | community | LMArena Leaderboard Dataset | |
| Hard Prompts English (style control)older version arena=text · category=hard_prompts_english · style_control=true | 1505.5 1500.1–1510.8 | community | LMArena Leaderboard Dataset | |
| French (style control)older version arena=text · category=french · style_control=true | 1504.0 1489.7–1518.3 | community | LMArena Leaderboard Dataset | |
| Japanese (style control)older version arena=text · category=japanese · style_control=true | 1503.2 1479.2–1527.1 | community | LMArena Leaderboard Dataset | |
| Industry Medicine And Healthcare (style control)older version arena=text · category=industry_medicine_and_healthcare · style_control=true | 1501.1 1492.3–1509.9 | community | LMArena Leaderboard Dataset | |
| Russian (style control)older version arena=text · category=russian · style_control=true | 1500.4 1492.8–1508.0 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1499.3 1494.5–1504.2 | community | LMArena Leaderboard Dataset | |
| Longer Query (style control)older version arena=text · category=longer_query · style_control=true | 1497.6 1492.4–1502.8 | community | LMArena Leaderboard Dataset | |
| Polish (style control)older version arena=text · category=polish · style_control=true | 1497.0 1482.1–1511.9 | community | LMArena Leaderboard Dataset | |
| Industry Legal And Government (style control)older version arena=text · category=industry_legal_and_government · style_control=true | 1495.6 1487.1–1504.2 | community | LMArena Leaderboard Dataset | |
| Multi Turn (style control)older version arena=text · category=multi_turn · style_control=true | 1492.6 1486.3–1499.0 | community | LMArena Leaderboard Dataset | |
| German (style control)older version arena=text · category=german · style_control=true | 1492.5 1474.5–1510.4 | community | LMArena Leaderboard Dataset | |
| Math (style control)older version arena=text · category=math · style_control=true | 1491.0 1481.3–1500.6 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1487.8 1483.1–1492.5 | community | LMArena Leaderboard Dataset | |
| Industry Mathematical (style control)older version arena=text · category=industry_mathematical · style_control=true | 1486.6 1476.4–1496.8 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1485.6 1482.0–1489.2 | community | LMArena Leaderboard Dataset | |
| Spanish (style control)older version arena=text · category=spanish · style_control=true | 1482.8 1469.1–1496.5 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1480.9 1475.6–1486.2 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1480.4 1475.8–1484.9 | community | LMArena Leaderboard Dataset | |
| Industry Writing And Literature And Language (style control)older version arena=text · category=industry_writing_and_literature_and_language · style_control=true | 1478.5 1472.8–1484.3 | community | LMArena Leaderboard Dataset | |
| Creative Writing (style control)older version arena=text · category=creative_writing · style_control=true | 1478.2 1471.5–1484.8 | community | LMArena Leaderboard Dataset | |
| Industry Business And Management And Financial Operations (style control)older version arena=text · category=industry_business_and_management_and_financial_operations · style_control=true | 1475.8 1469.7–1481.8 | community | LMArena Leaderboard Dataset | |
| Industry Entertainment And Sports And Media (style control)older version arena=text · category=industry_entertainment_and_sports_and_media · style_control=true | 1464.0 1457.8–1470.1 | community | LMArena Leaderboard Dataset | |
| Korean (style control)older version arena=text · category=korean · style_control=true | 1463.5 1445.0–1482.0 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond implementation=vals-ai | high effort | 95.5 % | independent | Vals AI |
| GPQA Diamond | 94.1 % | independent | Epoch AI Benchmarking Hub 2026-02-20 | |
| GPQA Diamond implementation=artificial-analysis | 94.1 % | independent | Artificial Analysis | |
| Humanity's Last Exam contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale | high effort | 46.4 % 44.5–48.4 | independent | Scale Labs |
| Humanity's Last Exam implementation=artificial-analysis | 44.7 % | independent | Artificial Analysis | |
| MMLU-Pro implementation=vals-ai | high effort | 91.0 % | independent | Vals AI |
language
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort | 85.4 % | independent | LiveBench |
long context instruction
| AA-LCR | 72.7 % | independent | Artificial Analysis | |
| IFBench | 77.1 % | independent | Artificial Analysis | |
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort | 79.1 % | independent | LiveBench |
| MultiChallenge | 71.4 % 69.6–73.1 | independent | Scale Labs |
multimodal
| MMMU implementation=vals-ai | high effort | 88.2 % | independent | Vals AI |
professional
| CaseLaw | high effort | 64.8 % | independent | Vals AI |
| CorpFin | high effort | 64.5 % | independent | Vals AI |
| LegalBench | high effort | 87.4 % | independent | Vals AI |
| MedQA | high effort | 96.4 % | independent | Vals AI |
| TaxEval | high effort | 72.9 % | independent | Vals AI |
reasoning math
| AIME implementation=vals-ai | high effort | 98.1 % | independent | Vals AI |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 98.0 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=public_eval · model_type=CoT | 97.2 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 88.1 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 77.1 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 1.0 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 1.0 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 0.5 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-3older version split=semi_private · model_type=CoT | 0.4 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=public_eval · model_type=CoT | 0.4 usd_per_task | independent | ARC Prize Leaderboard | |
| EnigmaEval contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions. | 19.8 % 17.5–22.0 | independent | Scale Labs | |
| FrontierMath | 36.9 % | independent | Epoch AI Benchmarking Hub 2026-02-19 | |
| FrontierMath Tier 4older version | 16.7 % | independent | Epoch AI Benchmarking Hub 2026-02-19 | |
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort | 91.0 % | independent | LiveBench |
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort | 84.0 % | independent | LiveBench |
| OTIS Mock AIME 2024–2025 | 95.6 % | independent | Epoch AI Benchmarking Hub 2026-02-20 |
Agent + model results
systems, not bare-model scores
| agent + model judy + Gemini 3.1 Pro Preview | Terminal-Bench 2.0 | 80.2 % | unverified | Terminal-Bench Leaderboard |
| agent + model terminus-3-3 + Gemini 3.1 Pro Preview | Terminal-Bench 2.0 | 74.8 % | unverified | Terminal-Bench Leaderboard |
| agent + model Artificial Analysis harness + Gemini 3.1 Pro Preview | Terminal-Bench 2.1 | 73.8 % | independent | Artificial Analysis |
| agent + model gemini-cli + Gemini 3.1 Pro Preview | Terminal-Bench 2.1 | 70.7 % | community | Terminal-Bench Leaderboard |
| agent + model terminus-2 + Gemini 3.1 Pro Preview | Terminal-Bench 2.1 | 70.3 % | community | Terminal-Bench Leaderboard |
| agent + model gemini-cli + Gemini 3.1 Pro Preview | Terminal-Bench 2.0 | 61.4 % | unverified | Terminal-Bench Leaderboard |
| agent + model gemini-cli + Gemini 3.1 Pro Preview | Terminal-Bench 2.0 | 59.4 % | unverified | Terminal-Bench Leaderboard |
| agent + model Artificial Analysis harness + Gemini 3.1 Pro Preview | Terminal-Bench Hard | 53.8 % | independent | Artificial Analysis |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.