Gemma 4 31B (medium) vs Gemma 4 26B A4B (medium) vs Gemini 3 PRO Preview (medium) benchmark comparisonGemma 4 26B A4B (medium) leads on Score with 6.6. Gemma 4 31B (medium) leads on Reliability with 10.0. Gemma 4 26B A4B (medium) has the lowest Total Cost at $0.089. Gemini 3 PRO Preview (medium) is fastest at 9.05s.
Recommended modelGemma 4 26B A4B (medium)It has the best score here (6.6), while costing about 3.1x less than the other models in this comparison.
A test is fully passed only if every run passed for that test.Wrong answer: 3No answer: 2Timed out: 2Invalid tool call: 1Response Time (avg)103.83sResponse Time (max)912.19sResponse Time (total)2180.47sA test is fully passed only if every run passed for that test.…
A test is fully passed only if every run passed for that test.API error: 4Wrong answer: 3Response Time (avg)9.05sResponse Time (max)26.24sResponse Time (total)90.53sA test is fully passed only if every run passed for that test.…
Attempt pass rate
68.2%Attempt pass rate = passed attempts / total attempts across runs.…
66.7%Attempt pass rate = passed attempts / total attempts across runs.…
63.6%Attempt pass rate = passed attempts / total attempts across runs.…
Flaky tests
2Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
2Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
Total Runs
66Total Runs…
66Total Runs…
63Total Runs…
Cost per result
1.044Shows the average cost per correct benchmark answer in cents (lower is better).…
0.643Shows the average cost per correct benchmark answer in cents (lower is better).…
1.406Shows the average cost per correct benchmark answer in cents (lower is better).…
Total Cost
$0.163Total Cost (Current Price)…
$0.089Total Cost (Current Price)…
$0.385Total Cost (Current Price)…
Input Price
$0.220 / 1MInput Price…
$0.070 / 1MInput Price…
$9.506 / 1MInput Price…
Output Price
$0.550 / 1MOutput Price…
$0.340 / 1MOutput Price…
$9.506 / 1MOutput Price…
Total Input Tokens
94,992Total Input Tokens…
77,550Total Input Tokens…
28,848Total Input Tokens…
Output Tokens
34,468Output Tokens…
28,036Output Tokens…
1,490Output Tokens…
Reasoning Tokens
223,278Reasoning Tokens…
219,491Reasoning Tokens…
10,102Reasoning Tokens…
Response Time (avg)
75.38sResponse Time (avg)…
103.83sResponse Time (avg)…
9.05sResponse Time (avg)…
Response Time (max)
437.40sResponse Time (max)…
912.19sResponse Time (max)…
26.24sResponse Time (max)…
Response Time (total)
1507.52sResponse Time (total)…
2180.47sResponse Time (total)…
90.53sResponse Time (total)…
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#110 Gemma 4 31B
medium
Cost
$0.002
Time
45.7s
Tokens
2,696 tok
#95 Gemma 4 26B A4B
medium
Invalid SVG
Cost
$0.000
Time
300.0s
Tokens
0 tok
#133 Gemini 3 PRO Preview
medium
No endpoints found for google/gemini-3-pro-preview.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)12.89sResponse Time (max)26.66sResponse Time (total)51.55sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)6.20sResponse Time (max)9.64sResponse Time (total)24.78sA test is fully passed only if every run passed for that test.…
6.20sResponse Time (avg)…
816Total Input Tokens…
1,142Output Tokens…
3,045Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)14.99sResponse Time (max)26.24sResponse Time (total)29.99sA test is fully passed only if every run passed for that test.…
4.3Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
5.8Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
22.2%Attempt pass rate = passed attempts / total attempts across runs.…
1Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Timed out: 2No answer: 1Response Time (avg)219.76sResponse Time (max)437.40sResponse Time (total)659.27sA test is fully passed only if every run passed for that test.…
2.9Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No answer: 2Timed out: 1Response Time (avg)272.54sResponse Time (max)369.32sResponse Time (total)817.61sA test is fully passed only if every run passed for that test.…
272.54sResponse Time (avg)…
5,062Total Input Tokens…
14,838Output Tokens…
44,567Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
3.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.API error: 3Response Time (avg)0msResponse Time (max)0msResponse Time (total)0msA test is fully passed only if every run passed for that test.…
2.9Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
5.8Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
16.7%Attempt pass rate = passed attempts / total attempts across runs.…
1Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.API error: 1Invalid tool call: 1Response Time (avg)433.11sResponse Time (max)433.11sResponse Time (total)433.11sA test is fully passed only if every run passed for that test.…
6.3Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
50.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Invalid tool call: 1Response Time (avg)492.87sResponse Time (max)912.19sResponse Time (total)985.73sA test is fully passed only if every run passed for that test.…
492.87sResponse Time (avg)…
54,390Total Input Tokens…
5,451Output Tokens…
132,113Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
1.5Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
5.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 1Response Time (avg)10.37sResponse Time (max)10.37sResponse Time (total)10.37sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)21.11sResponse Time (max)21.94sResponse Time (total)42.21sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)16.51sResponse Time (max)20.57sResponse Time (total)33.02sA test is fully passed only if every run passed for that test.…
16.51sResponse Time (avg)…
8,334Total Input Tokens…
1,567Output Tokens…
2,827Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)10.84sResponse Time (max)10.84sResponse Time (total)10.84sA test is fully passed only if every run passed for that test.…
7.7Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
66.7%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 1Response Time (avg)38.48sResponse Time (max)68.92sResponse Time (total)115.43sA test is fully passed only if every run passed for that test.…
2.9Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
4.4Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
22.2%Attempt pass rate = passed attempts / total attempts across runs.…
2Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 2Timed out: 1Response Time (avg)23.62sResponse Time (max)27.00sResponse Time (total)47.23sA test is fully passed only if every run passed for that test.…
23.62sResponse Time (avg)…
516Total Input Tokens…
2,469Output Tokens…
7,105Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
5.3Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
33.3%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 2Response Time (avg)7.01sResponse Time (max)7.01sResponse Time (total)7.01sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)9.57sResponse Time (max)9.57sResponse Time (total)9.57sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)29.76sResponse Time (max)29.76sResponse Time (total)29.76sA test is fully passed only if every run passed for that test.…
29.76sResponse Time (avg)…
567Total Input Tokens…
25Output Tokens…
5,075Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)9.34sResponse Time (max)9.34sResponse Time (total)9.34sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)12.76sResponse Time (max)17.53sResponse Time (total)25.52sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)17.54sResponse Time (max)21.25sResponse Time (total)35.08sA test is fully passed only if every run passed for that test.…
17.54sResponse Time (avg)…
777Total Input Tokens…
887Output Tokens…
4,470Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
9.8Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)3.26sResponse Time (max)3.26sResponse Time (total)3.26sA test is fully passed only if every run passed for that test.…
9.9Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)26.91sResponse Time (max)61.08sResponse Time (total)80.72sA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)5.79sResponse Time (max)6.85sResponse Time (total)17.36sA test is fully passed only if every run passed for that test.…
5.79sResponse Time (avg)…
801Total Input Tokens…
410Output Tokens…
2,128Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)3.88sResponse Time (max)4.23sResponse Time (total)7.77sA test is fully passed only if every run passed for that test.…
3.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.API error: 1Response Time (avg)0msResponse Time (max)0msResponse Time (total)0msA test is fully passed only if every run passed for that test.…
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)9.01sResponse Time (max)9.01sResponse Time (total)9.01sA test is fully passed only if every run passed for that test.…
9.01sResponse Time (avg)…
6,096Total Input Tokens…
450Output Tokens…
1,256Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
10.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
100.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.No failed answers.Response Time (avg)11.96sResponse Time (max)11.96sResponse Time (total)11.96sA test is fully passed only if every run passed for that test.…
3.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 1Response Time (avg)90.14sResponse Time (max)90.14sResponse Time (total)90.14sA test is fully passed only if every run passed for that test.…
3.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.Wrong answer: 1Response Time (avg)180.87sResponse Time (max)180.87sResponse Time (total)180.87sA test is fully passed only if every run passed for that test.…
180.87sResponse Time (avg)…
191Total Input Tokens…
797Output Tokens…
16,905Reasoning Tokens…
Gemini 3 PRO PreviewArchived model: this model is no longer updated or tested on new tests.
3.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
10.0Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
0.0%Attempt pass rate = passed attempts / total attempts across runs.…
0Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
A test is fully passed only if every run passed for that test.API error: 1Response Time (avg)0msResponse Time (max)0msResponse Time (total)0msA test is fully passed only if every run passed for that test.…