Benchmark Almanac

llava-v1.6-vicuna-13b

llava-v1.6-vicuna-13b ranks first on NaturalBench.

Released January 31st, 2024 Compare with others
22
Benchmarks with a score
1
Ranked first, out of five or more models
1
In the top three
Unknown lab Jan 2024

llava-v1.6-vicuna-13b

No Epoch Capabilities Index score yet
  • VisualPuzzles 25.5% #29 /29
  • MM-Vet 44.9% #103 /164
  • SEED-Bench-2-Plus 55.6% #92 /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 65.7% #117 of 178 Official leaderboard

Vision, audio and robotics

BenchmarkScoreRankSource
NaturalBench 19.2% #1 of 6 Official leaderboard
ScienceQA 73.6% #6 of 17 Claim in Falcon2-11B Technical Report, via Epoch AI
WildVision 33.9 #7 of 14 Official leaderboard
MIRB 29.4% #4 of 7 Official leaderboard
MMT-Bench 53% #9 of 15 Official leaderboard
MM-Vet 44.9% #103 of 164 Official leaderboard
MathVerse 15.6% #6 of 9 Official leaderboard
SEED-Bench-2-Plus 55.6% #92 of 136 Official leaderboard
VCR (Visual Caption Restoration) 18.1% #29 of 42 Official leaderboard
WildVision Arena 956 #7 of 10 Official leaderboard
Image2Struct 31.6 #13 of 18 Official leaderboard
MMBench 70% #112 of 155 Official leaderboard
CCBench (MMBench Chinese Culture) 28.8% #115 of 152 Official leaderboard
MMBench-EN v1.1 67.3% #136 of 178 Official leaderboard
MMStar 40.4% #141 of 173 Official leaderboard
A-Bench 64.3% #98 of 116 Official leaderboard
DynaMath 19.8% #10 of 11 Official leaderboard
MMMU 36.2% #124 of 136 Official leaderboard
MMMU-Pro 17.2% #55 of 57 Official leaderboard
VisualPuzzles 25.5% #29 of 29 Official leaderboard

Other

BenchmarkScoreRankSource
OpenVLM Leaderboard 47.7 #136 of 173 Official leaderboard