Capability breakdown
Intelligence over time
Provider & region options
Fastest loop: DeepInfra · US East (Virginia) · 2.4 s end to end.
| DeepInfra | US East (Virginia) | 246 tok/s | 67 ms | 2.4 s | $0.09 | – |
| DeepInfra | EU Central (Frankfurt) | 246 tok/s | 89 ms | 2.5 s | $0.09 | – |
| vLLM (self-host) | US East (Virginia) | 219 tok/s | 66 ms | 2.7 s | $0.08 | – |
| vLLM (self-host) | EU West (Ireland) | 219 tok/s | 90 ms | 2.7 s | $0.08 | – |
| Fireworks | US East (Virginia) | 310 tok/s | 59 ms | 2.7 s | $0.12 | – |
| Fireworks | US West (Oregon) | 310 tok/s | 59 ms | 2.7 s | $0.12 | – |
| Fireworks | EU West (Ireland) | 310 tok/s | 75 ms | 2.8 s | $0.12 | – |
| Overshoot | US East (Virginia) | 406 tok/s | 35 ms | 2.9 s | $0.12 | – |
| Overshoot | US West (Oregon) | 406 tok/s | 42 ms | 2.9 s | $0.12 | – |
| Overshoot | EU West (Ireland) | 406 tok/s | 51 ms | 2.9 s | $0.12 | – |
| Overshoot | EU Central (Frankfurt) | 406 tok/s | 51 ms | 2.9 s | $0.12 | – |
| Overshoot | Asia Pacific (Tokyo) | 406 tok/s | 68 ms | 2.9 s | $0.12 | – |
| Overshoot | Middle East (Dubai) | 406 tok/s | 83 ms | 2.9 s | $0.12 | – |
| Overshoot | Asia Pacific (Mumbai) | 406 tok/s | 87 ms | 2.9 s | $0.12 | – |
| Overshoot | South America (São Paulo) | 406 tok/s | 91 ms | 2.9 s | $0.12 | – |
Sample outputs
Illustrative samples shared across models, not verbatim output from Qwen3.6 35B A3B.
Scene description
“Describe what's happening in this camera frame.”
A forklift is reversing toward loading bay 3 while a worker in a hi-vis vest signals from the left. Two pallets remain unwrapped near the dock edge.
OCR + extraction
“Read the shipping label and return structured JSON.”
{ "tracking": "1Z A24 9W7 03 8421 990 4", "carrier": "UPS", "weight_kg": 12.4, "dest": "Portland, OR 97204" }
Grounding
“Point to every person not wearing a hard hat.”
2 detections, [x:0.41, y:0.62, w:0.08, h:0.19] conf 0.94 · [x:0.77, y:0.55, w:0.07, h:0.21] conf 0.88.
Related models
Where it ranks
- #3Claude Sonnet 4.682.7
- #3Gemini 3.1 Pro82.7
- #5Qwen3.6 35B A3B80.9
- #6Qwen3.6 27B77.0
- #7GPT-5.4 mini75.5
- #8Gemini 3 Flash75.3