Benchmark Almanac

Aria

Aria ranks first on 2 of the 25 benchmarks with five or more models scored.

Released September 30th, 2024 Compare with others
25
Benchmarks with a score
2
Ranked first, out of five or more models
6
In the top three
Unknown lab Sep 2024

Aria

No Epoch Capabilities Index score yet
  • MMLongBench-Doc 28.3% #16 /17
  • MM-Vet 56.9% #68 /164
  • SEED-Bench-2-Plus 66.4% #46 /136

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

BenchmarkScoreRankSource
MMBench-CN v1.1 76.6% #70 of 178 Official leaderboard

Math and science

BenchmarkScoreRankSource
PHYSICS 12.9% #12 of 20 Official leaderboard
PHYSICS-Verified 9.20% #18 of 19 Official leaderboard

Vision, audio and robotics

BenchmarkScoreRankSource
WildVision 74.1 #1 of 14 Official leaderboard
LongVideoBench 65% #2 of 23 Official leaderboard
MMBench-Video 1.85 #3 of 28 Official leaderboard
TempCompass 69.6% #2 of 13 Official leaderboard
TVBench 51% #1 of 6 Official leaderboard
TemporalBench 25% #2 of 10 Official leaderboard
MMStar 60.7% #58 of 173 Official leaderboard
CoreCognition 58.4% #40 of 119 Official leaderboard
SEED-Bench-2-Plus 66.4% #46 of 136 Official leaderboard
Video-MME 72.1% #9 of 26 Official leaderboard
MMBench 80% #57 of 155 Official leaderboard
MMBench-EN v1.1 78.2% #70 of 178 Official leaderboard
MM-Vet 56.9% #68 of 164 Official leaderboard
Video-MMMU 50.8% #10 of 24 Official leaderboard
VRBench 54.5% #8 of 19 Official leaderboard
CCBench (MMBench Chinese Culture) 48.4% #69 of 152 Official leaderboard
MMVU 49.3% #13 of 28 Official leaderboard
A-Bench 71.3% #62 of 116 Official leaderboard
MMMU 54.9% #74 of 136 Official leaderboard
MEGA-Bench 33.1% #12 of 22 Official leaderboard
MMLongBench-Doc 28.3% #16 of 17 Official leaderboard

Other

BenchmarkScoreRankSource
OpenVLM Leaderboard 64.0 #67 of 173 Official leaderboard