Gemini 1.0 Pro
Gemini 1.0 Pro ranks first on DecodingTrust.
Google DeepMind Released 13 Dec 2023 37
Benchmarks with a score
1
Ranked first, out of five or more models
3
In the top three
Google DeepMind Dec 2023 Gemini 1.0 Pro
Closed weights
117.1
#209 of 271
Epoch Capabilities Index
- GPQA Diamond 34% #189 /219
- Vectara Hallucination Leaderboard (HHEM) 7.70% #33 /148
- ThaiExam 47.7% #15 /24
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vectara Hallucination Leaderboard (HHEM) | 7.70% | #33 of 148 | Official leaderboard | |
| EQ-Bench v2 | 75.1 | #37 of 85 | Official leaderboard | |
| MMLU | 70% | #56 of 127 | Epoch AI | |
| CriticEval | 5.57 | #4 of 8 | Official leaderboard | |
| AlpacaEval 2.0 | 24.4% | #22 of 40 | Official leaderboard | |
| MMBench-CN v1.1 | 69.2% | #104 of 178 | Official leaderboard | |
| ThaiExam | 47.7% | #15 of 24 | Official leaderboard | |
| MixEval | 78.9% | #23 of 34 | Official leaderboard | |
| LMArena (Elo) | 1,237 | #266 of 333 | Official leaderboard | |
| AlpacaEval | 79.7% | #16 of 19 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 34% | #189 of 219 | Epoch AI | |
| MATH Level 5 | 11.2% | #87 of 94 | Epoch AI | |
| GSM1k | 79.8% | #14 of 15 | Official leaderboard | |
| OTIS Mock AIME 2024-2025 | 1.11% | #193 of 202 | Epoch AI | |
| DTBench | 9.77% | #174 of 178 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MBPP+ | 61.4% | #20 of 35 | Official leaderboard | |
| CodeEditorBench | 73.7% | #4 of 7 | Official leaderboard | |
| HumanEval+ | 55.5% | #30 of 46 | Official leaderboard | |
| SEAL Coding | 685 | #25 of 26 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SnakeBench | 35 | #10 of 183 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LegalBench | 70.3% | #115 of 138 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BLINK | 45.7% | #2 of 11 | Official leaderboard | |
| CCBench (MMBench Chinese Culture) | 52.5% | #60 of 152 | Official leaderboard | |
| MathVerse | 23.5% | #4 of 9 | Official leaderboard | |
| MMBench | 73.6% | #91 of 155 | Official leaderboard | |
| MMBench-EN v1.1 | 70.2% | #118 of 178 | Official leaderboard | |
| MME | 1496.6 | #11 of 16 | Official leaderboard | |
| MMMU | 47.9% | #95 of 136 | Official leaderboard | |
| MathVista | 45.2% | #32 of 43 | Official leaderboard | |
| CharXiv | 22.8% | #63 of 77 | Official leaderboard | |
| MMStar | 38.6% | #148 of 173 | Official leaderboard | |
| MATH-Vision (MathVision) | 17.7% | #93 of 105 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OR-Bench | 9.70% | #2 of 28 | Official leaderboard | |
| DecodingTrust | 80.6 | #1 of 5 | Official leaderboard | |
| AIR-Bench 2024 | 58.2% | #55 of 82 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 56.1 | #104 of 173 | Official leaderboard | |
| HELM Lite | 42.2 | #48 of 71 | Official leaderboard |
Google DeepMind's models over time
Google DeepMind's models in color, against every other model on the Epoch Capabilities Index, by release date.