InternVL3-8B
InternVL3-8B ranks first on MM-Vet.
36
Benchmarks with a score
1
Ranked first, out of five or more models
2
In the top three
Shanghai AI Lab Apr 2025 InternVL3-8B
Open weights Small
No Epoch Capabilities Index score yet
- Video SimpleQA 23.5% #15 /27
- OCR-Reasoning 11.5% #14 /15
- VideoReasonBench 0.60% #11 /14
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 | 81.9% | #33 of 178 | Official leaderboard | |
| Video SimpleQA | 23.5% | #15 of 27 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| PhyX | 6.50% | #8 of 10 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SArena | 0.81 | #13 of 17 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MM-Vet | 82.8% | #1 of 164 | Official leaderboard | |
| CCBench (MMBench Chinese Culture) | 77.8% | #13 of 152 | Official leaderboard | |
| MMStar | 68.7% | #18 of 173 | Official leaderboard | |
| SEED-Bench-2-Plus | 69.4% | #18 of 136 | Official leaderboard | |
| MMBench | 83.6% | #25 of 155 | Official leaderboard | |
| A-Bench | 76.3% | #24 of 116 | Official leaderboard | |
| MMBench-EN v1.1 | 82.3% | #37 of 178 | Official leaderboard | |
| We-Math | 39.5% | #6 of 21 | Official leaderboard | |
| CharXiv (DQ) | 73.6% | #21 of 67 | Official leaderboard | |
| CharXiv | 37.6% | #26 of 77 | Official leaderboard | |
| MMBench-Video | 1.69 | #10 of 28 | Official leaderboard | |
| MMMU | 62.7% | #50 of 136 | Official leaderboard | |
| Video-Holmes | 32.3% | #2 of 5 | Official leaderboard | |
| ERQA | 37.1% | #8 of 17 | Official leaderboard | |
| VideoEval-Pro | 24.7% | #5 of 10 | Official leaderboard | |
| Ego3D-Bench | 8.00 | #6 of 11 | Official leaderboard | |
| CrossVid | 25.6% | #8 of 13 | Official leaderboard | |
| SpaCE-10 | 40% | #23 of 36 | Official leaderboard | |
| EmbodiedBench EB-Manipulation | 11.5% | #12 of 17 | Official leaderboard | |
| EmbodiedBench EB-Navigation | 29.3% | #12 of 17 | Official leaderboard | |
| MATH-Vision (MathVision) | 29% | #76 of 105 | Official leaderboard | |
| Video-Oasis | 31.6% | #3 of 4 | Official leaderboard | |
| EmbodiedBench EB-Habitat | 24.3% | #13 of 17 | Official leaderboard | |
| SpatialViz-Bench | 30.3% | #11 of 14 | Official leaderboard | |
| VideoReasonBench | 0.60% | #11 of 14 | Official leaderboard | |
| EmbodiedBench EB-ALFRED | 10.3% | #14 of 17 | Official leaderboard | |
| MME-CC | 11.4% | #10 of 12 | Official leaderboard | |
| MM-HELIX | 8.10% | #15 of 17 | Official leaderboard | |
| MMSI-Bench | 25.7% | #25 of 28 | Official leaderboard | |
| MORSE-500 | 7.80% | #10 of 11 | Official leaderboard | |
| OCR-Reasoning | 11.5% | #14 of 15 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 73.6 | #16 of 173 | Official leaderboard |