HN 표시: AI Benchy – AI 벤치마크 및 비교

AI BENCHY AI Benchmark Leaderboard Last updated at: 2026-03-30 Models Evaluated: 78 78/80 Filter models No models match the current search and filters. | Rank | Model | Score Summarizes broad quality across our full private benchmark suite, so ranking reflects consistent performance. | Company | Total Cost | Response Time (avg) Response Time (avg) | Tests Correct Shows how many tests are fully passed (all runs pass). | |---|---|---|---|---|---|---| | #1🥇 #1 | Gemini 3 Flash Previewmedium | 10.0… | $0.166… | 11.39s… | A test is fully passed only if every run passed for that test. No failed answers. Response Time (avg)11.39s Response Time (max)50.16s Response Time (total)113.86s … | | | |||||| | #2🥈 #2 | Gemini 3.1 Pro Previewmedium | 9.6… | $0.522… | 15.56s… | A test is fully passed only if every run passed for that test. Wrong answer: 1 Response Time (avg)15.56s Response Time (max)40.61s Response Time (total)155.64s … | | | |||||| | #3🥉 #3 | Gemini 3 Flash Previewlow | 8.7… | $0.081… | 5.95s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Response Time (avg)5.95s Response Time (max)14.72s Response Time (total)101.19s … | | | |||||| | #4#4 | Gemini 3 PRO Previewmedium | 8.7… | $0.197… | 9.06s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Response Time (avg)9.06s Response Time (max)26.24s Response Time (total)90.58s … | | | |||||| | #5#5 | Seed-2.0-Litemedium | 8.5… | Bytedance Seed | $0.105… | 27.78s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 2 Response Time (avg)27.78s Response Time (max)168.71s Response Time (total)472.24s … | | |||||| | #6#6 | Qwen3.6 Plus Previewmedium | 8.5… | Qwen | $0.000… | 13.94s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 1 Response Time (avg)13.94s Response Time (max)43.55s Response Time (total)237.01s … | | |||||| | #7#7 | GPT-5.3-Codexmedium | 8.5… | OpenAI | $0.544… | 15.76s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 2 Response Time (avg)15.76s Response Time (max)100.93s Response Time (total)267.97s … | | |||||| | #8#8 | Gemini 3.1 Flash Lite PreviewhighArchived model: this model is no longer updated or tested on new tests. | 8.4… | $2.310… | 68.83s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 1 Response Time (avg)68.83s Response Time (max)280.52s Response Time (total)1101.32s … | | | |||||| | #9#9 | Qwen3.5 Plus 2026-02-15medium | 8.4… | Qwen | $0.189… | 39.13s… | A test is fully passed only if every run passed for that test. Timed out: 2 Wrong answer: 2 Response Time (avg)39.13s Response Time (max)81.20s Response Time (total)391.29s … | | |||||| | #10#10 | Qwen3.5-122B-A10Bmedium | 8.4… | Qwen | $0.505… | 29.05s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Timed out: 1 Response Time (avg)29.05s Response Time (max)119.29s Response Time (total)493.86s … | | |||||| | #11#11 | Qwen3.5-27Bmedium | 8.3… | Qwen | $0.467… | 52.01s… | A test is fully passed only if every run passed for that test. Did not follow instructions: 2 Extra formatting: 1 Timed out: 1 Wrong answer: 1 Response Time (avg)52.01s Response Time (max)163.96s Response Time (total)884.10s … | | |||||| | #12#12 | GLM 5medium | 8.3… | Z.ai | $0.108… | 17.15s… | A test is fully passed only if every run passed for that test. Wrong answer: 2 Did not follow instructions: 1 No answer: 1 Timed out: 1 Response Time (avg)17.15s Response Time (max)28.96s Response Time (total)154.32s … | | |||||| | #13#13 | DeepSeek V3.2medium | 8.2… | DeepSeek | $0.026… | 38.49s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 1 Timed out: 1 Response Time (avg)38.49s Response Time (max)93.11s Response Time (total)654.41s … | | |||||| | #14#14 | Gemini 2.5 Flashmedium | 8.1… | $0.292… | 11.88s… | A test is fully passed only if every run passed for that test. Wrong answer: 4 Did not follow instructions: 1 Response Time (avg)11.88s Response Time (max)95.48s Response Time (total)201.89s … | | | |||||| | #15#15 | Gemini 3.1 Flash Lite Previewmedium | 8.1… | $0.050… | 3.70s… | A test is fully passed only if every run passed for that test. Wrong answer: 4 Did not follow instructions: 1 Response Time (avg)3.70s Response Time (max)14.93s Response Time (total)62.97s … | | | |||||| | #16#16 | GPT-5.4medium | 8.1… | OpenAI | $0.794… | 18.95s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 2 Response Time (avg)18.95s Response Time (max)100.41s Response Time (total)322.23s … | | |||||| | #17#17 | GLM 5 Turbomedium | 8.0… | Z.ai | $0.166… | 17.98s… | A test is fully passed only if every run passed for that test. Wrong answer: 3 Did not follow instructions: 2 Timed out: