Skip to main content
dbb.Latest results

Dog Breed Bench.

Which models can identify dog breeds the best?

Version 1

Tested Sep 29, 2026

100 photos · 1 attempt per photo

Model comparison ranked by first-choice breed accuracy. Accuracy uses a fixed 0–100% scale. Response time and cost bars scale to the largest value in their column.
ModelRanked by accuracyAccuracyCorrect first-choice breedResponse timeMedian · secondsCost / 1,000 callsEstimated · USD
Gemini 3.8 FlashGoogle · mediumAccuracy
93.0%
Median response
5.63 s
Cost / 1,000 calls
$4.28
Gemini 3.1 Pro PreviewGoogle · medium1 of 100 calls failedAccuracy
92.0%
Median response
4.08 s
Cost / 1,000 calls
$7.32
Claude Sonnet 5.5Anthropic · medium3 of 100 calls failedAccuracy
91.0%
Median response
3.27 s
Cost / 1,000 calls
$10.77
GPT-6 SolOpenAI · mediumAccuracy
84.0%
Median response
8.20 s
Cost / 1,000 calls
$9.46
GPT-6.1 SolOpenAI · mediumAccuracy
83.0%
Median response
6.48 s
Cost / 1,000 calls
$8.63
GPT-5.6 SolOpenAI · mediumAccuracy
83.0%
Median response
4.71 s
Cost / 1,000 calls
$16.58
GPT-6 LunaOpenAI · mediumAccuracy
75.0%
Median response
5.89 s
Cost / 1,000 calls
$0.53
GPT-6 LunaOpenAI · max15 of 100 calls failedAccuracy
69.0%
Median response
24.43 s
Cost / 1,000 calls
$1.32
GPT-5.6 LunaOpenAI · mediumAccuracy
68.0%
Median response
4.43 s
Cost / 1,000 calls
$0.95
GPT-5.6 LunaOpenAI · max10 of 100 calls failedAccuracy
63.0%
Median response
12.11 s
Cost / 1,000 calls
$2.52
Claude Opus 5.5Anthropic · mediumRun incomplete — 4/100 calls; 3 failedUnavailableProvider access unavailable (HTTP 429).
Time and cost bars use separate scales.

Accuracy vs. cost

First-choice accuracy (%)

First-choice breed accuracy versus estimated costGemini 3.8 Flash · medium: 93.0% first-choice accuracy, $4.28 estimated per 1,000 calls; 95% confidence interval 88.0% to 97.0%Gemini 3.1 Pro Preview · medium: 92.0% first-choice accuracy, $7.32 estimated per 1,000 calls; 95% confidence interval 86.0% to 97.0%Claude Sonnet 5.5 · medium: 91.0% first-choice accuracy, $10.77 estimated per 1,000 calls; 95% confidence interval 85.0% to 96.0%GPT-6 Sol · medium: 84.0% first-choice accuracy, $9.46 estimated per 1,000 calls; 95% confidence interval 77.0% to 91.0%GPT-6.1 Sol · medium: 83.0% first-choice accuracy, $8.63 estimated per 1,000 calls; 95% confidence interval 75.0% to 90.0%GPT-5.6 Sol · medium: 83.0% first-choice accuracy, $16.58 estimated per 1,000 calls; 95% confidence interval 75.0% to 90.0%GPT-6 Luna · medium: 75.0% first-choice accuracy, $0.53 estimated per 1,000 calls; 95% confidence interval 66.0% to 83.0%GPT-6 Luna · max: 69.0% first-choice accuracy, $1.32 estimated per 1,000 calls; 95% confidence interval 60.0% to 78.0%GPT-5.6 Luna · medium: 68.0% first-choice accuracy, $0.95 estimated per 1,000 calls; 95% confidence interval 59.0% to 77.0%GPT-5.6 Luna · max: 63.0% first-choice accuracy, $2.52 estimated per 1,000 calls; 95% confidence interval 54.0% to 73.0%First-choice breed accuracy versus estimated costGemini 3.8 Flash · medium: 93.0% first-choice accuracy, $4.28 estimated per 1,000 calls; 95% confidence interval 88.0% to 97.0%Gemini 3.1 Pro Preview · medium: 92.0% first-choice accuracy, $7.32 estimated per 1,000 calls; 95% confidence interval 86.0% to 97.0%Claude Sonnet 5.5 · medium: 91.0% first-choice accuracy, $10.77 estimated per 1,000 calls; 95% confidence interval 85.0% to 96.0%GPT-6 Sol · medium: 84.0% first-choice accuracy, $9.46 estimated per 1,000 calls; 95% confidence interval 77.0% to 91.0%GPT-6.1 Sol · medium: 83.0% first-choice accuracy, $8.63 estimated per 1,000 calls; 95% confidence interval 75.0% to 90.0%GPT-5.6 Sol · medium: 83.0% first-choice accuracy, $16.58 estimated per 1,000 calls; 95% confidence interval 75.0% to 90.0%GPT-6 Luna · medium: 75.0% first-choice accuracy, $0.53 estimated per 1,000 calls; 95% confidence interval 66.0% to 83.0%GPT-6 Luna · max: 69.0% first-choice accuracy, $1.32 estimated per 1,000 calls; 95% confidence interval 60.0% to 78.0%GPT-5.6 Luna · medium: 68.0% first-choice accuracy, $0.95 estimated per 1,000 calls; 95% confidence interval 59.0% to 77.0%GPT-5.6 Luna · max: 63.0% first-choice accuracy, $2.52 estimated per 1,000 calls; 95% confidence interval 54.0% to 73.0%

Estimated USD per 1,000 calls · logarithmic scale

  1. Gemini 3.8 FlashGoogle · medium93.0% · $4.28 / 1,000 calls
  2. Gemini 3.1 Pro PreviewGoogle · medium92.0% · $7.32 / 1,000 calls
  3. Claude Sonnet 5.5Anthropic · medium91.0% · $10.77 / 1,000 calls
  4. GPT-6 SolOpenAI · medium84.0% · $9.46 / 1,000 calls
  5. GPT-6.1 SolOpenAI · medium83.0% · $8.63 / 1,000 calls
  6. GPT-5.6 SolOpenAI · medium83.0% · $16.58 / 1,000 calls
  7. GPT-6 LunaOpenAI · medium75.0% · $0.53 / 1,000 calls
  8. GPT-6 LunaOpenAI · max69.0% · $1.32 / 1,000 calls
  9. GPT-5.6 LunaOpenAI · medium68.0% · $0.95 / 1,000 calls
  10. GPT-5.6 LunaOpenAI · max63.0% · $2.52 / 1,000 calls
Whiskers show 95% confidence intervals. Only complete runs with at least half of calls successful are plotted.

Claude Sonnet 5.5’s saved scores include 10 responses that exceeded the prompt’s note-length limit. Strict response-format validation needs re-evaluation; these archived scores have not been changed.

Gemini 3.1 Pro Preview’s saved scores include two responses that exceeded the prompt’s note-length limit by one and three characters. These scores follow the archived scoring rules; strict response-format compliance is reported separately in the repository audit.

Accuracy by dog group

Correct / attempts

0%100%
ModelSporting11 photosHound12 photosWorking21 photosTerrier13 photosToy16 photosNon-Sporting12 photosHerding13 photos
Gemini 3.8 FlashGoogle · medium
90.9%10 / 11 correct
100.0%12 / 12 correct
100.0%21 / 21 correct
84.6%11 / 13 correct
81.3%13 / 16 correct
100.0%12 / 12 correct
92.3%12 / 13 correct
Gemini 3.1 Pro PreviewGoogle · medium
90.9%10 / 11 correct
91.7%11 / 12 correct
95.2%20 / 21 correct
92.3%12 / 13 correct
81.3%13 / 16 correct
100.0%12 / 12 correct
92.3%12 / 13 correct
Claude Sonnet 5.5Anthropic · medium
90.9%10 / 11 correct
100.0%12 / 12 correct
81.0%17 / 21 correct
84.6%11 / 13 correct
100.0%16 / 16 correct
100.0%12 / 12 correct
84.6%11 / 13 correct
GPT-6 SolOpenAI · medium
81.8%9 / 11 correct
91.7%11 / 12 correct
85.7%18 / 21 correct
61.5%8 / 13 correct
87.5%14 / 16 correct
100.0%12 / 12 correct
76.9%10 / 13 correct
GPT-6.1 SolOpenAI · medium
81.8%9 / 11 correct
91.7%11 / 12 correct
81.0%17 / 21 correct
69.2%9 / 13 correct
87.5%14 / 16 correct
100.0%12 / 12 correct
69.2%9 / 13 correct
GPT-5.6 SolOpenAI · medium
90.9%10 / 11 correct
83.3%10 / 12 correct
76.2%16 / 21 correct
76.9%10 / 13 correct
87.5%14 / 16 correct
100.0%12 / 12 correct
69.2%9 / 13 correct
GPT-6 LunaOpenAI · medium
81.8%9 / 11 correct
66.7%8 / 12 correct
81.0%17 / 21 correct
61.5%8 / 13 correct
68.8%11 / 16 correct
100.0%12 / 12 correct
61.5%8 / 13 correct
GPT-6 LunaOpenAI · max
72.7%8 / 11 correct
75.0%9 / 12 correct
81.0%17 / 21 correct
38.5%5 / 13 correct
50.0%8 / 16 correct
100.0%12 / 12 correct
61.5%8 / 13 correct
GPT-5.6 LunaOpenAI · medium
36.4%4 / 11 correct
66.7%8 / 12 correct
61.9%13 / 21 correct
84.6%11 / 13 correct
75.0%12 / 16 correct
91.7%11 / 12 correct
53.8%7 / 13 correct
GPT-5.6 LunaOpenAI · max
54.5%6 / 11 correct
58.3%7 / 12 correct
61.9%13 / 21 correct
69.2%9 / 13 correct
56.3%9 / 16 correct
91.7%11 / 12 correct
46.2%6 / 13 correct
Claude Opus 5.5Anthropic · mediumRun incomplete — 4/100 calls; 3 failed
Unavailable
Unavailable
Unavailable
Unavailable
Unavailable
Unavailable
Unavailable

AKC groups, small samples. Unavailable: incomplete run or over half of calls failed in the run or group.

2 photos outside these groups count toward overall accuracy: Dutch Shepherd (Miscellaneous Class), 1 photo; Australian Kelpie (Foundation Stock Service), 1 photo.

Photos & test conditions

Photos
One photo per breed. Identical JPEGs for every model, 600–1,536 pixels per edge. Lighting and framing vary.
Labels
Checked against source descriptions and photos, without independent expert review. No mixed breeds or non-dog photos.
Settings
Same prompt, one pass, 60-second timeout, no retries. Reasoning varies by provider. First-choice matches count; failed calls count as incorrect.

Small score differences may be noise. Costs are estimated from token usage. Request pacing varied, so response times may not be comparable.