Results
500 photos across 25 dog breeds. 7 models.
| Model | First match | Top three | Median latency |
|---|---|---|---|
| GPT-6.1 TerraMedium reasoning | First match | Top three | Median latency4.66 s |
| GPT-6 AstraMedium reasoning | First match | Top three | Median latency9.43 s |
| GPT-6.1 SolMedium reasoning | First match | Top three | Median latency7.99 s |
| GPT-5 MiniMedium reasoning | First match | Top three | Median latency12.77 s |
| GPT-4o MiniNo reasoning setting | First match | Top three | Median latency2.01 s |
| GPT-6 Luna FastMedium reasoning | First match | Top three | Median latency3.69 s |
| GPT-6 LunaMedium reasoning | First match | Top three | Median latency5.26 s |
Median latency is measured per request. Serving conditions varied between runs.
Accuracy by breed
Each breed has 20 photos.
Average gives each of the 7 models equal weight.
| Breed | Average | Terra | Astra | Sol | 5 Mini | 4o Mini | Luna Fast | Luna |
|---|---|---|---|---|---|---|---|---|
| Staffordshire Bull Terrier | 50.7%7 models | 60%12/20 | 75%15/20 | 80%16/20 | 55%11/20 | 40%8/20 | 30%6/20 | 15%3/20 |
| American Pit Bull Terrier | 47.9%7 models | 45%9/20 | 45%9/20 | 50%10/20 | 75%15/20 | 65%13/20 | 40%8/20 | 15%3/20 |
| English Setter | 71.4%7 models | 90%18/20 | 85%17/20 | 85%17/20 | 90%18/20 | 75%15/20 | 30%6/20 | 45%9/20 |
| American Bulldog | 61.4%7 models | 95%19/20 | 70%14/20 | 65%13/20 | 40%8/20 | 70%14/20 | 45%9/20 | 45%9/20 |
| English Cocker Spaniel | 62.9%7 models | 75%15/20 | 75%15/20 | 80%16/20 | 55%11/20 | 25%5/20 | 70%14/20 | 60%12/20 |
| Miniature Pinscher | 76.4%7 models | 50%10/20 | 85%17/20 | 90%18/20 | 100%20/20 | 100%20/20 | 55%11/20 | 55%11/20 |
| Scottish Terrier | 56.4%7 models | 65%13/20 | 75%15/20 | 70%14/20 | 45%9/20 | 60%12/20 | 45%9/20 | 35%7/20 |
| Havanese | 54.3%7 models | 75%15/20 | 45%9/20 | 45%9/20 | 70%14/20 | 55%11/20 | 50%10/20 | 40%8/20 |
| Keeshond | 87.9%7 models | 95%19/20 | 95%19/20 | 95%19/20 | 95%19/20 | 95%19/20 | 60%12/20 | 80%16/20 |
| Japanese Chin | 83.6%7 models | 90%18/20 | 90%18/20 | 85%17/20 | 90%18/20 | 60%12/20 | 85%17/20 | 85%17/20 |
| Leonberger | 87.9%7 models | 100%20/20 | 95%19/20 | 100%20/20 | 85%17/20 | 85%17/20 | 70%14/20 | 80%16/20 |
| Soft-coated Wheaten Terrier | 84.3%7 models | 85%17/20 | 95%19/20 | 95%19/20 | 95%19/20 | 80%16/20 | 65%13/20 | 75%15/20 |
| Chihuahua | 79.3%7 models | 85%17/20 | 75%15/20 | 65%13/20 | 75%15/20 | 90%18/20 | 85%17/20 | 80%16/20 |
| Newfoundland | 81.4%7 models | 80%16/20 | 80%16/20 | 80%16/20 | 75%15/20 | 70%14/20 | 90%18/20 | 95%19/20 |
| Samoyed | 87.1%7 models | 90%18/20 | 95%19/20 | 95%19/20 | 90%18/20 | 95%19/20 | 75%15/20 | 70%14/20 |
| Shiba Inu | 87.9%7 models | 90%18/20 | 85%17/20 | 85%17/20 | 100%20/20 | 100%20/20 | 75%15/20 | 80%16/20 |
| Basset Hound | 91.4%7 models | 95%19/20 | 100%20/20 | 90%18/20 | 90%18/20 | 80%16/20 | 90%18/20 | 95%19/20 |
| Great Pyrenees | 91.4%7 models | 95%19/20 | 90%18/20 | 90%18/20 | 100%20/20 | 80%16/20 | 95%19/20 | 90%18/20 |
| Boxer | 83.6%7 models | 85%17/20 | 85%17/20 | 85%17/20 | 85%17/20 | 90%18/20 | 75%15/20 | 80%16/20 |
| Yorkshire Terrier | 87.1%7 models | 95%19/20 | 90%18/20 | 85%17/20 | 85%17/20 | 95%19/20 | 80%16/20 | 80%16/20 |
| Beagle | 93.6%7 models | 90%18/20 | 90%18/20 | 95%19/20 | 95%19/20 | 100%20/20 | 95%19/20 | 90%18/20 |
| German Shorthaired Pointer | 85.0%7 models | 85%17/20 | 90%18/20 | 90%18/20 | 80%16/20 | 80%16/20 | 85%17/20 | 85%17/20 |
| Pomeranian | 94.3%7 models | 100%20/20 | 90%18/20 | 90%18/20 | 95%19/20 | 100%20/20 | 95%19/20 | 90%18/20 |
| Pug | 99.3%7 models | 95%19/20 | 100%20/20 | 100%20/20 | 100%20/20 | 100%20/20 | 100%20/20 | 100%20/20 |
| Saint Bernard | 85.7%7 models | 85%17/20 | 90%18/20 | 85%17/20 | 85%17/20 | 85%17/20 | 85%17/20 | 85%17/20 |
Methodology
Each model evaluated the same 500 photos once. A match requires a valid response with the reference breed first or among the top three. All attempts count.
Dataset credit: Oxford-IIIT Pet dataset (Parkhi, Vedaldi, Zisserman, and Jawahar, 2012), CC BY-SA 4.0. Photo copyrights remain with their owners.