👁️
LEADERBOARD
Vision Models
Image understanding, classification, and visual reasoning models
15tools ranked
Vision Models Rankings
Ranked by overall ToolRoute Score across all benchmark dimensions
| Rank | Tool Name | ToolRoute Score | Output | Reliability | Efficiency | Cost | Trust | Stars |
|---|---|---|---|---|---|---|---|---|
| 🥇 | SAM 2 | 7.0 | 6.9 | 6.6 | 6.8 | 7.2 | 8.5 | 19,046 |
| 🥈 | GPT-4 VisionOfficial | 6.8 | 7.1 | 6.9 | 6.5 | 5.0 | 9.0 | 30,637 |
| 🥉 | Claude 3 VisionOfficial | 6.8 | 7.0 | 7.0 | 6.6 | 5.0 | 9.0 | 3,340 |
| #4 | Moondream | 6.8 | 6.3 | 6.2 | 7.2 | 7.2 | 8.7 | 9,616 |
| #5 | RoboflowOfficial | 6.8 | 6.7 | 6.6 | 6.5 | 5.7 | 9.5 | 38,259 |
| #6 | PixtralOfficial | 6.7 | 6.7 | 6.6 | 6.8 | 5.5 | 9.3 | 888 |
| #7 | Florence-2 | 6.7 | 6.7 | 6.5 | 6.9 | 7.2 | 6.6 | 8,000 |
| #8 | InternVL | 6.7 | 6.8 | 6.5 | 6.7 | 7.2 | 6.5 | 10,003 |
| #9 | Grounding DINO | 6.6 | 6.7 | 6.4 | 6.6 | 7.2 | 6.5 | 10,057 |
| #10 | CogVLM | 6.6 | 6.7 | 6.4 | 6.6 | 7.2 | 6.5 | 6,737 |
| #11 | LLaVA | 6.6 | 6.6 | 6.4 | 6.7 | 7.2 | 6.5 | 24,736 |
| #12 | Qwen-VL | 6.6 | 6.7 | 6.5 | 6.8 | 7.1 | 6.0 | 6,639 |
| #13 | Azure Computer VisionOfficial | 6.5 | 6.4 | 6.7 | 6.6 | 5.5 | 8.0 | 318 |
| #14 | Gemini VisionOfficial | 6.4 | 6.9 | 6.8 | 6.9 | 5.3 | 4.5 | 2,291 |
| #15 | AWS RekognitionOfficial | 6.3 | 6.5 | 6.8 | 6.6 | 5.5 | 5.0 | 7,626 |
💡
Why SAM 2 is #1
SAM 2 leads GPT-4 Vision by +2.2 in Cost, and also wins in Efficiency.
Output Quality
6.9
vs 7.1
Reliability
6.6
vs 6.9
Efficiency
6.8
vs 6.5
Cost
7.2
vs 5.0
Trust
8.5
vs 9.0
Score Guide:9+ Exceptional8+ Excellent7+ Good6+ Fair<6 Below Avg
Contribute Benchmark Data
Help improve these rankings by submitting real-world telemetry. Contributors earn routing credits for every data point.