gpt-oss-120b (medium) leads on average score with 5.9 vs 5.8. Qwen3.8 Flash Next has the lower benchmark cost at ~$0.011 vs $0.030. Qwen3.8 Flash Next is faster at 5.71s vs 31.58s, with pass rates of 47.8% vs 43.5%.
Last updated at:
2026-10-05
Compared models
Rank
#252
Total Output Tokens
118,831
Response Time (avg)
31.58s
Total Cost
$0.030
Candidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
Rank
#258
Total Output Tokens
9,077
Response Time (avg)
5.71s
Total Cost
~$0.011Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.1095 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.010823.
Recommended modelQwen3.8 Flash Next
Its score stays close to the best score here (5.8 vs 5.9), while costing about 2.7x less than gpt-oss-120b (medium).
Qwen3.8 Flash NextQwen3.8 Flash NextnoneCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.Release: 2026-10-04
Qwen3.8 Flash NextQwen3.8 Flash NextnoneCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.Release: 2026-10-04
Score
5.9Summarizes 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
#252
#258
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
8.1Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
9.0Consistency 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: 10Did not follow instructions: 3Invalid tool call: 1Response Time (avg)31.58sResponse Time (max)203.90sResponse Time (total)536.78sA 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: 14No answer: 1Response Time (avg)5.71sResponse Time (max)65.25sResponse Time (total)131.36sA test is fully passed only if every run passed for that test.…
Attempt pass rate
47.8%Attempt pass rate = passed attempts / total attempts across runs.…
43.5%Attempt pass rate = passed attempts / total attempts across runs.…
Flaky tests
5Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
3Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
Total Runs
69Total Runs…
69Total Runs…
Cost per result
0.335
~0.136Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.1095 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.010823.
Total Cost
$0.030
~$0.011Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.1095 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.010823.
Input Price
$0.037 / 1MInput Price…
N/AInput Price…
Output Price
$0.170 / 1MOutput Price…
N/AOutput Price…
Cache Read Price
N/ACache Read Price…
N/ACache Read Price…
Cache Write Price
N/ACache Write Price…
N/ACache Write Price…
Total Input Tokens
285,293Total Input Tokens…
266,004Total Input Tokens…
Output Tokens
29,565Output Tokens…
9,077Output Tokens…
Reasoning Tokens
89,266Reasoning Tokens…
0Reasoning Tokens…
Response Time (avg)
31.58sResponse Time (avg)…
5.71sResponse Time (avg)…
Response Time (max)
203.90sResponse Time (max)…
65.25sResponse Time (max)…
Response Time (total)
536.78sResponse Time (total)…
131.36sResponse Time (total)…
Parameters
117B total (5.1B active)
180B total (6B active)
Availability
Open source
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.
#252 gpt-oss-120b
medium
Cost
$0.001
Time
26.7s
Tokens
555 tok
#258 Qwen3.8 Flash Next
noneCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
Cost
~$0.001Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.0076 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.000755.
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.Wrong answer: 1Response Time (avg)203.90sResponse Time (max)203.90sResponse Time (total)203.90sA test is fully passed only if every run passed for that test.…
203.90sResponse Time (avg)…
176,526Total Input Tokens…
187Output Tokens…
20,362Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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.Wrong answer: 1Response Time (avg)65.25sResponse Time (max)65.25sResponse Time (total)65.25sA test is fully passed only if every run passed for that test.…
6.7Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
9.9Consistency 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.Did not follow instructions: 1Wrong answer: 1Response Time (avg)10.21sResponse Time (max)19.76sResponse Time (total)20.43sA test is fully passed only if every run passed for that test.…
10.21sResponse Time (avg)…
1,314Total Input Tokens…
3,518Output Tokens…
2,177Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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: 2Response Time (avg)1.70sResponse Time (max)4.03sResponse Time (total)6.82sA 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.…
7.0Consistency 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)38.37sResponse Time (max)68.16sResponse Time (total)115.10sA test is fully passed only if every run passed for that test.…
38.37sResponse Time (avg)…
7,782Total Input Tokens…
3,365Output Tokens…
11,973Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
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).…
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)1.64sResponse Time (max)2.70sResponse Time (total)4.91sA test is fully passed only if every run passed for that test.…
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.Invalid tool call: 1Response Time (avg)23.96sResponse Time (max)31.18sResponse Time (total)47.91sA test is fully passed only if every run passed for that test.…
23.96sResponse Time (avg)…
81,198Total Input Tokens…
10,453Output Tokens…
22,883Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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: 1Wrong answer: 1Response Time (avg)11.98sResponse Time (max)16.02sResponse Time (total)23.95sA test is fully passed only if every run passed for that test.…
6.4Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
5.9Consistency 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)1.98sResponse Time (max)1.98sResponse Time (total)1.98sA test is fully passed only if every run passed for that test.…
1.98sResponse Time (avg)…
7,476Total Input Tokens…
241Output Tokens…
1,114Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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)1.60sResponse Time (max)1.71sResponse Time (total)3.19sA 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: 3Response Time (avg)27.55sResponse Time (max)50.92sResponse Time (total)55.10sA test is fully passed only if every run passed for that test.…
27.55sResponse Time (avg)…
1,286Total Input Tokens…
6,390Output Tokens…
21,466Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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.Wrong answer: 3Response Time (avg)5.38sResponse Time (max)15.16sResponse Time (total)16.15sA 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.…
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)7.90sResponse Time (max)7.90sResponse Time (total)7.90sA test is fully passed only if every run passed for that test.…
7.90sResponse Time (avg)…
659Total Input Tokens…
107Output Tokens…
387Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
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)1.09sResponse Time (max)1.09sResponse Time (total)1.09sA 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)7.63sResponse Time (max)7.63sResponse Time (total)7.63sA test is fully passed only if every run passed for that test.…
7.63sResponse Time (avg)…
1,036Total Input Tokens…
126Output Tokens…
1,799Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
7.1Summarizes 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).…
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.Wrong answer: 1Response Time (avg)612msResponse Time (max)637msResponse Time (total)1.22sA 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.…
7.2Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
44.4%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: 1Wrong answer: 1Response Time (avg)21.71sResponse Time (max)32.40sResponse Time (total)43.41sA test is fully passed only if every run passed for that test.…
21.71sResponse Time (avg)…
1,190Total Input Tokens…
1,790Output Tokens…
2,264Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · NVIDIA GeForce RTX 3090.
6.3Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
7.8Consistency 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)1.30sResponse Time (max)2.37sResponse Time (total)3.89sA 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)6.91sResponse Time (max)6.91sResponse Time (total)6.91sA test is fully passed only if every run passed for that test.…
6.91sResponse Time (avg)…
6,514Total Input Tokens…
287Output Tokens…
1,083Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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)4.35sResponse Time (max)4.35sResponse Time (total)4.35sA 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)26.51sResponse Time (max)26.51sResponse Time (total)26.51sA test is fully passed only if every run passed for that test.…
26.51sResponse Time (avg)…
312Total Input Tokens…
3,101Output Tokens…
3,758Reasoning Tokens…
Qwen3.8 Flash NextCandidate responses were generated on local hardware. OpenRouter was used only for judging. qwen3.8-flash-next-iq3_s · IQ3_S · Strata 0.1.39 · 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.Wrong answer: 1Response Time (avg)540msResponse Time (max)540msResponse Time (total)540msA test is fully passed only if every run passed for that test.…