Benchmark Almanac

InternVL2-Llama3-76B

InternVL2-Llama3-76B places in the top three on 2 benchmarks.

Shanghai AI Lab Released July 15th, 2024 Family InternVL2 The lab's model card Compare with others
21
Benchmarks with a score
0
Ranked first, out of five or more models
2
In the top three
Shanghai AI Lab Jul 2024

InternVL2-Llama3-76B

Open weights Flagship

No Epoch Capabilities Index score yet
  • HR-Bench 66.9% #3 /8
  • EMMA 25.1% #7 /7
  • MM-Vet 64.4% #41 /164

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 85.5% #13 of 178 Official leaderboard
SimpleVQA 35.7% #9 of 13 Official leaderboard

Coding

BenchmarkScoreRankSource
ChartMimic 58.5 #2 of 14 Official leaderboard

Vision, audio and robotics

BenchmarkScoreRankSource
CCBench (MMBench Chinese Culture) 81.2% #6 of 152 Official leaderboard
MMBench 86.5% #12 of 155 Official leaderboard
MMBench-EN v1.1 85.5% #14 of 178 Official leaderboard
VCR (Visual Caption Restoration) 48.3% #5 of 42 Official leaderboard
SEED-Bench-2-Plus 70% #17 of 136 Official leaderboard
MMStar 67.1% #26 of 173 Official leaderboard
MM-Vet 64.4% #41 of 164 Official leaderboard
CharXiv (DQ) 75.2% #17 of 67 Official leaderboard
CharXiv 38.9% #24 of 77 Official leaderboard
We-Math 36.9% #7 of 21 Official leaderboard
A-Bench 73.9% #42 of 116 Official leaderboard
MEGA-Bench 37.7% #8 of 22 Official leaderboard
HR-Bench 66.9% #3 of 8 Official leaderboard
MMMU 58.3% #64 of 136 Official leaderboard
MTVQA-Test 22% #9 of 18 Official leaderboard
MMMU-Pro 40% #34 of 57 Official leaderboard
EMMA 25.1% #7 of 7 Official leaderboard

Other

BenchmarkScoreRankSource
OpenVLM Leaderboard 71.0 #29 of 173 Official leaderboard