The average score is effectively tied at 7.2 vs 7.2. Qwen3.8 27B (medium) has the lower benchmark cost at ~$0.073 vs $3.787. Claude Opus 4.8 (low) is faster at 17.13s vs 40.09s, with pass rates of 78.3% vs 66.7%.
Last updated at:
2026-11-13
Compared models
Rank
#211
Total Output Tokens
69,489
Response Time (avg)
17.13s
Total Cost
$3.787
Candidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
Rank
#209
Total Output Tokens
170,696
Response Time (avg)
40.09s
Total Cost
~$0.073Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.7349 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.072660.
Recommended modelQwen3.8 27B (medium)
It has the best score here (7.2), while costing about 52.1x less than Claude Opus 4.8 (low).
Qwen3.8 27BQwen3.8 27BmediumCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.Release: 2026-08-14Free Available
Qwen3.8 27BQwen3.8 27BmediumCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.Release: 2026-08-14Free Available
Score
7.2Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.2Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
Rank
#211
#209
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.…
9.7First-attempt success score: 10.0 means no retryable target API or rate-limit failures before successful calls; tracked failures lower the score.…
Consistency
8.6Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
9.4Consistency 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: 5Extra formatting: 1No answer: 1Response Time (avg)17.13sResponse Time (max)127.97sResponse Time (total)394.01sA test is fully passed only if every run passed for that test.…
78.3%Attempt pass rate = passed attempts / total attempts across runs.…
66.7%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).…
2Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
Total Runs
69Total Runs…
69Total Runs…
Cost per result
23.669
~0.520Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.7349 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.072660.
Total Cost
$3.787
~$0.073Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.7349 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.072660.
Input Price
$5.000 / 1MInput Price…
N/AInput Price…
Output Price
$25.000 / 1MOutput Price…
N/AOutput Price…
Cache Read Price
$0.500 / 1MCache Read Price…
N/ACache Read Price…
Cache Write Price
$6.250 / 1MCache Write Price…
N/ACache Write Price…
Total Input Tokens
409,947Total Input Tokens…
277,164Total Input Tokens…
Output Tokens
46,060Output Tokens…
1,913Output Tokens…
Reasoning Tokens
23,429Reasoning Tokens…
168,783Reasoning Tokens…
Response Time (avg)
17.13sResponse Time (avg)…
40.09sResponse Time (avg)…
Response Time (max)
127.97sResponse Time (max)…
214.63sResponse Time (max)…
Response Time (total)
394.01sResponse Time (total)…
881.89sResponse Time (total)…
Parameters
~5T total (~500B active)
27.3B
Availability
Closed
Weights available
Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.
Model generation showcase
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#211 Claude Opus 4.8
low
Cost
$0.031
Time
14.1s
Tokens
1,345 tok
#209 Qwen3.8 27B
mediumCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
Cost
~$0.002Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.0121 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.001193.
4.7Summarizes 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).…
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.Wrong answer: 1Response Time (avg)110.68sResponse Time (max)110.68sResponse Time (total)110.68sA test is fully passed only if every run passed for that test.…
110.68sResponse Time (avg)…
253,425Total Input Tokens…
3,004Output Tokens…
14,248Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
4.7Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
1.6Consistency 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)187.85sResponse Time (max)187.85sResponse Time (total)187.85sA 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)3.30sResponse Time (max)4.34sResponse Time (total)13.18sA test is fully passed only if every run passed for that test.…
3.30sResponse Time (avg)…
834Total Input Tokens…
793Output Tokens…
371Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)2.93sResponse Time (max)5.18sResponse Time (total)11.70sA test is fully passed only if every run passed for that test.…
6.6Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
4.6Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
77.8%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.Extra formatting: 1Wrong answer: 1Response Time (avg)7.58sResponse Time (max)12.19sResponse Time (total)22.74sA test is fully passed only if every run passed for that test.…
7.58sResponse Time (avg)…
10,590Total Input Tokens…
3,637Output Tokens…
809Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)40.50sResponse Time (max)72.14sResponse Time (total)121.51sA 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)36.87sResponse Time (max)52.90sResponse Time (total)73.74sA test is fully passed only if every run passed for that test.…
36.87sResponse Time (avg)…
119,079Total Input Tokens…
13,586Output Tokens…
2,867Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
8.7Summarizes 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.Invalid tool call: 1Response Time (avg)124.48sResponse Time (max)214.63sResponse Time (total)248.96sA 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.…
5.8Consistency 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)2.27sResponse Time (max)2.50sResponse Time (total)4.54sA test is fully passed only if every run passed for that test.…
2.27sResponse Time (avg)…
10,503Total Input Tokens…
310Output Tokens…
0Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
6.5Summarizes 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).…
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)8.76sResponse Time (max)10.36sResponse Time (total)17.53sA test is fully passed only if every run passed for that test.…
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)46.54sResponse Time (max)127.97sResponse Time (total)139.63sA test is fully passed only if every run passed for that test.…
46.54sResponse Time (avg)…
972Total Input Tokens…
23,226Output Tokens…
4,472Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)76.98sResponse Time (max)144.07sResponse Time (total)230.93sA 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)2.55sResponse Time (max)2.55sResponse Time (total)2.55sA test is fully passed only if every run passed for that test.…
2.55sResponse Time (avg)…
708Total Input Tokens…
231Output Tokens…
0Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)9.20sResponse Time (max)9.20sResponse Time (total)9.20sA test is fully passed only if every run passed for that test.…
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)2.78sResponse Time (max)3.78sResponse Time (total)5.56sA test is fully passed only if every run passed for that test.…
2.78sResponse Time (avg)…
909Total Input Tokens…
111Output Tokens…
221Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)2.54sResponse Time (max)2.67sResponse Time (total)5.07sA 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)3.01sResponse Time (max)3.31sResponse Time (total)9.04sA test is fully passed only if every run passed for that test.…
3.01sResponse Time (avg)…
894Total Input Tokens…
592Output Tokens…
184Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
8.3Summarizes 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).…
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.Did not follow instructions: 1Response Time (avg)12.62sResponse Time (max)29.62sResponse Time (total)37.85sA 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)6.85sResponse Time (max)6.85sResponse Time (total)6.85sA test is fully passed only if every run passed for that test.…
6.85sResponse Time (avg)…
11,775Total Input Tokens…
370Output Tokens…
35Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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.…
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)5.48sResponse Time (max)5.48sResponse Time (total)5.48sA test is fully passed only if every run passed for that test.…
5.48sResponse Time (avg)…
258Total Input Tokens…
200Output Tokens…
222Reasoning Tokens…
Qwen3.8 27BCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8:latest · Q4_K_M · Ollama 0.32.12 · NVIDIA GeForce RTX 3090.
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)11.29sResponse Time (max)11.29sResponse Time (total)11.29sA test is fully passed only if every run passed for that test.…