Pixtral 12B
Pixtral 12B.
Mistral AI Released 17 Sep 2024 18
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Mistral AI Sep 2024 Pixtral 12B
Open weights
No Epoch Capabilities Index score yet
- OCRBench v2 38.4% #44 /65
- HumanEval-V 6.30% #12 /24
- A-Bench 73.1% #50 /116
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 | |
|---|---|---|---|---|
| MMBench-CN v1.1 | 70.1% | #98 of 178 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval-V | 6.30% | #12 of 24 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SnakeBench | 17 | #153 of 183 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CharXiv | 42.4% | #19 of 77 | Official leaderboard | |
| A-Bench | 73.1% | #50 of 116 | Official leaderboard | |
| MMBench | 77.9% | #70 of 155 | Official leaderboard | |
| MMBench-EN v1.1 | 75.3% | #89 of 178 | Official leaderboard | |
| MEGA-Bench | 33.7% | #11 of 22 | Official leaderboard | |
| MMStar | 54.5% | #88 of 173 | Official leaderboard | |
| MathVista | 58% | #22 of 43 | Official leaderboard | |
| CCBench (MMBench Chinese Culture) | 37.6% | #95 of 152 | Official leaderboard | |
| Image2Struct | 41.2 | #12 of 18 | Official leaderboard | |
| OCRBench v2 | 38.4% | #44 of 65 | Official leaderboard | |
| VHELM | 29.6 | #34 of 41 | Official leaderboard | |
| VISTA | 26% | #50 of 54 | Official leaderboard | |
| GeoBench | 37% | #46 of 48 | Official leaderboard | |
| MMVU | 32.3% | #28 of 28 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 61.0 | #78 of 173 | Official leaderboard |