Nemotron 3 Ultra (medium) vs Qwen3.8 Flash Next (low)
The average score is effectively tied at 8.0 vs 8.0. Qwen3.8 Flash Next (low) has the lower benchmark cost at ~$0.045 vs $0.674. Qwen3.8 Flash Next (low) is faster at 23.66s vs 40.33s, with pass rates of 68.1% vs 79.7%.
Last updated at:
2026-10-05
Compared models
Rank
#93
Total Output Tokens
203,522
Response Time (avg)
40.33s
Total Cost
$0.674
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
#96
Total Output Tokens
135,295
Response Time (avg)
23.66s
Total Cost
~$0.045Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.4535 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.044834.
Recommended modelQwen3.8 Flash Next (low)
It has the best score here (8.0), while costing about 15.0x less than Nemotron 3 Ultra (medium).
Qwen3.8 Flash NextQwen3.8 Flash NextlowCandidate 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 NextlowCandidate 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
8.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
8.0Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
Rank
#93
#96
Reliability
9.8First-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
8.9Consistency score reflects run-to-run stability (10 = very consistent, even if consistently wrong).…
8.0Consistency 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: 7API error: 1Invalid tool call: 1Response Time (avg)40.33sResponse Time (max)392.56sResponse Time (total)927.65sA 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: 6Did not follow instructions: 1Invalid tool call: 1Response Time (avg)23.66sResponse Time (max)166.72sResponse Time (total)544.16sA test is fully passed only if every run passed for that test.…
Attempt pass rate
68.1%Attempt pass rate = passed attempts / total attempts across runs.…
79.7%Attempt pass rate = passed attempts / total attempts across runs.…
Flaky tests
3Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
6Flaky tests had mixed outcomes across runs (at least one pass and one fail).…
Total Runs
69Total Runs…
69Total Runs…
Cost per result
0.000
~0.299Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.4535 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.044834.
Total Cost
$0.674
~$0.045Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.4535 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.044834.
Input Price
$0.500 / 1MInput Price…
N/AInput Price…
Output Price
$2.200 / 1MOutput Price…
N/AOutput Price…
Cache Read Price
$0.100 / 1MCache Read Price…
N/ACache Read Price…
Cache Write Price
N/ACache Write Price…
N/ACache Write Price…
Total Input Tokens
482,754Total Input Tokens…
342,163Total Input Tokens…
Output Tokens
62,455Output Tokens…
1,518Output Tokens…
Reasoning Tokens
141,067Reasoning Tokens…
133,777Reasoning Tokens…
Response Time (avg)
40.33sResponse Time (avg)…
23.66sResponse Time (avg)…
Response Time (max)
392.56sResponse Time (max)…
166.72sResponse Time (max)…
Response Time (total)
927.65sResponse Time (total)…
544.16sResponse Time (total)…
Parameters
550B total (55B active)
180B total (6B active)
Availability
Weights available
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.
#93 Nemotron 3 Ultra
medium
Reached the allocated time limit (300 seconds) without receiving showcase output.
Cost
$0.000
Time
300.0s
Tokens
0 tok
#96 Qwen3.8 Flash Next
lowCandidate 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
~$0.002Estimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.0131 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.001295.
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)188.06sResponse Time (max)188.06sResponse Time (total)188.06sA test is fully passed only if every run passed for that test.…
188.06sResponse Time (avg)…
249,339Total Input Tokens…
8,071Output Tokens…
11,046Reasoning 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)137.50sResponse Time (max)137.50sResponse Time (total)137.50sA 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)8.62sResponse Time (max)16.86sResponse Time (total)34.49sA test is fully passed only if every run passed for that test.…
8.62sResponse Time (avg)…
780Total Input Tokens…
835Output Tokens…
1,485Reasoning 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)2.03sResponse Time (max)2.72sResponse Time (total)8.13sA 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.…
7.4Consistency 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)26.53sResponse Time (max)31.91sResponse Time (total)79.58sA test is fully passed only if every run passed for that test.…
26.53sResponse Time (avg)…
7,686Total Input Tokens…
2,854Output Tokens…
17,725Reasoning 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)29.46sResponse Time (max)49.90sResponse Time (total)88.39sA 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.Invalid tool call: 1Response Time (avg)218.25sResponse Time (max)392.56sResponse Time (total)436.49sA test is fully passed only if every run passed for that test.…
218.25sResponse Time (avg)…
204,249Total Input Tokens…
40,954Output Tokens…
78,561Reasoning 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.
6.9Summarizes 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.Invalid tool call: 1Response Time (avg)98.91sResponse Time (max)166.72sResponse Time (total)197.82sA 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)5.68sResponse Time (max)7.94sResponse Time (total)11.36sA test is fully passed only if every run passed for that test.…
5.68sResponse Time (avg)…
7,989Total Input Tokens…
473Output Tokens…
1,285Reasoning 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)6.20sResponse Time (max)6.23sResponse Time (total)12.40sA test is fully passed only if every run passed for that test.…
3.6Summarizes 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).…
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.Wrong answer: 3Response Time (avg)35.22sResponse Time (max)49.56sResponse Time (total)105.65sA test is fully passed only if every run passed for that test.…
35.22sResponse Time (avg)…
785Total Input Tokens…
7,637Output Tokens…
18,208Reasoning 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.
4.1Summarizes 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).…
44.5%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)23.34sResponse Time (max)42.75sResponse Time (total)70.03sA test is fully passed only if every run passed for that test.…
3.7Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance.…
9.5Consistency 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)2.52sResponse Time (max)2.52sResponse Time (total)2.52sA test is fully passed only if every run passed for that test.…
2.52sResponse Time (avg)…
360Total Input Tokens…
70Output Tokens…
235Reasoning 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.
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)5.00sResponse Time (max)5.00sResponse Time (total)5.00sA 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)6.35sResponse Time (max)9.38sResponse Time (total)12.69sA test is fully passed only if every run passed for that test.…
6.35sResponse Time (avg)…
765Total Input Tokens…
182Output Tokens…
1,243Reasoning 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.
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)2.32sResponse Time (max)3.23sResponse Time (total)4.64sA test is fully passed only if every run passed for that test.…
5.5Summarizes 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).…
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)3.54sResponse Time (max)6.03sResponse Time (total)10.62sA test is fully passed only if every run passed for that test.…
3.54sResponse Time (avg)…
792Total Input Tokens…
771Output Tokens…
2,055Reasoning 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.
8.7Summarizes 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).…
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.Did not follow instructions: 1Response Time (avg)1.72sResponse Time (max)2.26sResponse Time (total)5.15sA 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)7.72sResponse Time (max)7.72sResponse Time (total)7.72sA test is fully passed only if every run passed for that test.…
7.72sResponse Time (avg)…
9,781Total Input Tokens…
304Output Tokens…
984Reasoning 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)5.85sResponse Time (max)5.85sResponse Time (total)5.85sA 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)38.47sResponse Time (max)38.47sResponse Time (total)38.47sA test is fully passed only if every run passed for that test.…
38.47sResponse Time (avg)…
228Total Input Tokens…
304Output Tokens…
8,240Reasoning 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)9.26sResponse Time (max)9.26sResponse Time (total)9.26sA test is fully passed only if every run passed for that test.…