Gemini 3 Flash Preview vs Grok 4.20 Beta benchmark comparison: Gemini 3 Flash Preview leads on average score with 9.8 vs 5.8. Grok 4.20 Beta has the lower benchmark cost at $0.087 vs $0.667. Grok 4.20 Beta is faster at 1.19s vs 18.64s, with pass rates of 98.4% vs 37.0%.
Recommended model: Grok 4.20 Beta - It offers the best overall trade-off: a competitive score (5.8), lower cost than Gemini 3 Flash Preview, and balanced response time.
Grok 4.20 BetaGrok 4.20 BetanoneArchived model: this model is no longer updated or tested on new tests.Release: 2026-03-12
Score
9.8Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
5.8Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
Rank
#1
#107
Reliability
10.0First-attempt success score: 10.0 means no retryable target API or rate-limit failures before successful calls; tracked failures lower the score.…
N/AFirst-attempt success score: 10.0 means no retryable target API or rate-limit failures before successful calls; tracked failures lower the score.…
Consistency
9.7Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
9.6Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
Tests Correct
A test is fully passed only if every run passed for that test.Wrong answer: 1Response Time (avg)18.64sResponse Time (max)117.26sResponse Time (total)391.35sA 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.Wrong answer: 10Did not follow instructions: 1Invalid tool call: 1Response Time (avg)1.19sResponse Time (max)6.48sResponse Time (total)21.43sA test is fully passed only if every run passed for that test.…
Attempt pass rate
98.4%Attempt pass rate = passed attempts / total attempts across runs.…
37.0%Attempt pass rate = passed attempts / total attempts across runs.…
Flaky tests
1Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
1Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
Total Runs
63Total Runs…
52Total Runs…
Cost per result
3.335Shows the average cost per correct benchmark answer in cents (lower is better).…
1.510Shows the average cost per correct benchmark answer in cents (lower is better).…
Total Cost
$0.667Total Cost (Current Price)…
$0.087Total Cost (Current Price)…
Input Price
$0.500 / 1MInput Price…
$2.041 / 1MInput Price…
Output Price
$3.000 / 1MOutput Price…
$2.041 / 1MOutput Price…
Total Input Tokens
37,017Total Input Tokens…
40,597Total Input Tokens…
Output Tokens
2,006Output Tokens…
1,657Output Tokens…
Reasoning Tokens
214,153Reasoning Tokens…
0Reasoning Tokens…
Response Time (avg)
18.64sResponse Time (avg)…
1.19sResponse Time (avg)…
Response Time (max)
117.26sResponse Time (max)…
6.48sResponse Time (max)…
Response Time (total)
391.35sResponse Time (total)…
21.43sResponse Time (total)…
Generation showcase
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
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)5.73sResponse Time (total)15.53sA test is fully passed only if every run passed for that test.…
3.88sResponse Time (avg)…
494Total Input Tokens…
330Output Tokens…
3,216Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.
4.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
8.4Consistency 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.Wrong answer: 4Response Time (avg)597msResponse Time (max)866msResponse Time (total)2.39sA test is fully passed only if every run passed for that test.…
8.6Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.6Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
88.9%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.Wrong answer: 1Response Time (avg)84.40sResponse Time (max)117.26sResponse Time (total)253.21sA test is fully passed only if every run passed for that test.…
84.40sResponse Time (avg)…
8,122Total Input Tokens…
462Output Tokens…
161,084Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.
5.5Summarizes 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)1.14sResponse Time (max)1.14sResponse Time (total)1.14sA 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)22.42sResponse Time (max)22.42sResponse Time (total)22.42sA test is fully passed only if every run passed for that test.…
22.42sResponse Time (avg)…
12,873Total Input Tokens…
351Output Tokens…
10,485Reasoning Tokens…
Grok 4.20 BetaArchived 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.Invalid tool call: 1Response Time (avg)6.48sResponse Time (max)6.48sResponse Time (total)6.48sA 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)5.43sResponse Time (max)6.18sResponse Time (total)10.86sA test is fully passed only if every run passed for that test.…
5.43sResponse Time (avg)…
7,548Total Input Tokens…
279Output Tokens…
4,893Reasoning Tokens…
Grok 4.20 BetaArchived 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)601msResponse Time (max)634msResponse Time (total)1.20sA 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)15.27sResponse Time (max)34.09sResponse Time (total)45.80sA test is fully passed only if every run passed for that test.…
15.27sResponse Time (avg)…
633Total Input Tokens…
12Output Tokens…
21,684Reasoning Tokens…
Grok 4.20 BetaArchived 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.Wrong answer: 3Response Time (avg)611msResponse Time (max)616msResponse Time (total)1.83sA 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)5.19sResponse Time (max)5.19sResponse Time (total)5.19sA test is fully passed only if every run passed for that test.…
5.19sResponse Time (avg)…
486Total Input Tokens…
72Output Tokens…
1,905Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.
5.0Summarizes 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.Did not follow instructions: 1Response Time (avg)541msResponse Time (max)541msResponse Time (total)541msA 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)4.04sResponse Time (max)4.70sResponse Time (total)8.08sA test is fully passed only if every run passed for that test.…
4.04sResponse Time (avg)…
615Total Input Tokens…
72Output Tokens…
2,709Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.
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.Wrong answer: 1Response Time (avg)649msResponse Time (max)952msResponse Time (total)1.30sA 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)4.05sResponse Time (max)5.64sResponse Time (total)12.15sA test is fully passed only if every run passed for that test.…
4.05sResponse Time (avg)…
558Total Input Tokens…
183Output Tokens…
4,365Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.
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)586msResponse Time (max)813msResponse Time (total)1.76sA 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)12.60sResponse Time (max)12.60sResponse Time (total)12.60sA test is fully passed only if every run passed for that test.…
12.60sResponse Time (avg)…
5,532Total Input Tokens…
234Output Tokens…
1,487Reasoning Tokens…
Grok 4.20 BetaArchived 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)4.79sResponse Time (max)4.79sResponse Time (total)4.79sA 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)5.50sResponse Time (max)5.50sResponse Time (total)5.50sA test is fully passed only if every run passed for that test.…
5.50sResponse Time (avg)…
156Total Input Tokens…
11Output Tokens…
2,325Reasoning Tokens…
Grok 4.20 BetaArchived model: this model is no longer updated or tested on new tests.