GPT-6 Luna (high) leads on average score with 7.7 vs 7.6. GPT-6 Luna (high) has the lower benchmark cost at $0.074 vs ~$0.089. GPT-6 Luna (high) is faster at 20.37s vs 46.60s, with pass rates of 73.9% vs 68.1%.
Last updated at:
2026-10-01
Compared models
Rank
#111
Total Output Tokens
101,587
Response Time (avg)
20.37s
Total Cost
$0.074
Candidate responses were generated on local hardware. OpenRouter was used only for judging.
Rank
#117
Total Output Tokens
209,143
Response Time (avg)
46.60s
Total Cost
~$0.089Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.8932 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.088310.
Recommended modelGPT-6 Luna (high)
It has the best score here (7.7), while responding about 2.3x faster than Qwen3.8 27B (low).
Qwen3.8 27BQwen3.8 27BlowCandidate responses were generated on local hardware. OpenRouter was used only for judging.Release: 2026-08-14Free Available
Score
7.7Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.6Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
Rank
#111
#117
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.…
10.0First-attempt success score: 10.0 means no retryable target API or rate-limit failures before successful calls; tracked failures lower the score.…
Consistency
8.7Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
9.7Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
Attempts
69/69
69/69
Tests Correct
A test is fully passed only if every run passed for that test.Wrong answer: 6Did not follow instructions: 1Invalid tool call: 1Response Time (avg)20.37sResponse Time (max)106.93sResponse Time (total)468.52sA 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: 5No answer: 2Did not follow instructions: 1Response Time (avg)46.60sResponse Time (max)376.04sResponse Time (total)1071.83sA test is fully passed only if every run passed for that test.…
Attempt pass rate
73.9%Attempt pass rate = passed attempts / total attempts across runs.…
68.1%Attempt pass rate = passed attempts / total attempts across runs.…
Flaky tests
4Flaky 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
69Total Runs…
69Total Runs…
Cost per result
0.492
~0.589Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.8932 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.088310.
Total Cost
$0.074
~$0.089Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.8932 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.088310.
Input Price
$0.100 / 1MInput Price…
N/AInput Price…
Output Price
$0.500 / 1MOutput Price…
N/AOutput Price…
Total Input Tokens
229,986Total Input Tokens…
315,290Total Input Tokens…
Output Tokens
6,037Output Tokens…
1,607Output Tokens…
Reasoning Tokens
95,550Reasoning Tokens…
207,536Reasoning Tokens…
Response Time (avg)
20.37sResponse Time (avg)…
46.60sResponse Time (avg)…
Response Time (max)
106.93sResponse Time (max)…
376.04sResponse Time (max)…
Response Time (total)
468.52sResponse Time (total)…
1071.83sResponse Time (total)…
Parameters
~400B total (~17B active)
27.3B
Availability
Closed
Weights available
Model generation showcase
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#111 GPT-6 Luna
high
Cost
$0.003
Time
49.2s
Tokens
5,231 tok
#117 Qwen3.8 27B
lowCandidate responses were generated on local hardware. OpenRouter was used only for judging.
Cost
~$0.003Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.0258 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.002549.
5.2Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
3.2Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
33.3%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.Invalid tool call: 1Response Time (avg)82.78sResponse Time (max)82.78sResponse Time (total)82.78sA test is fully passed only if every run passed for that test.…
82.78sResponse Time (avg)…
149,545Total Input Tokens…
1,276Output Tokens…
8,906Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
6.1Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
3.1Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
66.7%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: 1Response Time (avg)211.32sResponse Time (max)211.32sResponse Time (total)211.32sA 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)2.64sResponse Time (max)3.26sResponse Time (total)10.56sA test is fully passed only if every run passed for that test.…
2.64sResponse Time (avg)…
606Total Input Tokens…
159Output Tokens…
1,358Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)3.02sResponse Time (max)5.68sResponse Time (total)12.07sA test is fully passed only if every run passed for that test.…
7.8Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.4Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
77.8%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)24.69sResponse Time (max)33.57sResponse Time (total)74.07sA test is fully passed only if every run passed for that test.…
24.69sResponse Time (avg)…
7,302Total Input Tokens…
483Output Tokens…
15,691Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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.No answer: 1Response Time (avg)140.92sResponse Time (max)376.04sResponse Time (total)422.75sA 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)21.53sResponse Time (max)28.77sResponse Time (total)43.07sA test is fully passed only if every run passed for that test.…
21.53sResponse Time (avg)…
57,233Total Input Tokens…
3,111Output Tokens…
6,658Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)90.63sResponse Time (max)143.71sResponse Time (total)181.27sA 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)24.50sResponse Time (max)27.15sResponse Time (total)48.99sA test is fully passed only if every run passed for that test.…
24.50sResponse Time (avg)…
7,140Total Input Tokens…
234Output Tokens…
352Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)8.03sResponse Time (max)9.09sResponse Time (total)16.06sA test is fully passed only if every run passed for that test.…
5.9Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.2Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
55.6%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: 2Response Time (avg)51.14sResponse Time (max)106.93sResponse Time (total)153.41sA test is fully passed only if every run passed for that test.…
51.14sResponse Time (avg)…
732Total Input Tokens…
70Output Tokens…
55,170Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)60.78sResponse Time (max)120.94sResponse Time (total)182.34sA test is fully passed only if every run passed for that test.…
5.1Summarizes 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)17.40sResponse Time (max)17.40sResponse Time (total)17.40sA test is fully passed only if every run passed for that test.…
17.40sResponse Time (avg)…
477Total Input Tokens…
96Output Tokens…
373Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)5.37sResponse Time (max)5.37sResponse Time (total)5.37sA test is fully passed only if every run passed for that test.…
8.4Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
6.9Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
83.3%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.Did not follow instructions: 1Response Time (avg)1.92sResponse Time (max)2.04sResponse Time (total)3.84sA test is fully passed only if every run passed for that test.…
1.92sResponse Time (avg)…
660Total Input Tokens…
91Output Tokens…
420Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)2.43sResponse Time (max)2.87sResponse Time (total)4.86sA test is fully passed only if every run passed for that test.…
7.8Summarizes 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).…
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)7.51sResponse Time (max)13.60sResponse Time (total)22.53sA test is fully passed only if every run passed for that test.…
7.51sResponse Time (avg)…
642Total Input Tokens…
231Output Tokens…
4,988Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)4.91sResponse Time (max)8.81sResponse Time (total)14.74sA 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)6.96sResponse Time (max)6.96sResponse Time (total)6.96sA test is fully passed only if every run passed for that test.…
6.96sResponse Time (avg)…
5,454Total Input Tokens…
256Output Tokens…
261Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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)8.42sResponse Time (max)8.42sResponse Time (total)8.42sA 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)4.90sResponse Time (max)4.90sResponse Time (total)4.90sA test is fully passed only if every run passed for that test.…
4.90sResponse Time (avg)…
195Total Input Tokens…
30Output Tokens…
1,373Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging.
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.No answer: 1Response Time (avg)12.64sResponse Time (max)12.64sResponse Time (total)12.64sA test is fully passed only if every run passed for that test.…