Compare / Qwen3 Next 80B A3B Thinking vs gpt-oss-20b

Qwen3 Next 80B A3B Thinking vs gpt-oss-20b

19 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.

Qwen3 Next 80B A3B Thinking
Alibaba · profile
gpt-oss-20b
OpenAI · profile
Results compared
19
Shared sources
3
BenchmarkMetricQwen3 Next 80B A3B Thinkinggpt-oss-20b
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,3671,323
Arena (formerly LMArena)reportedCreative writingArena rating 1,3261,240
Arena (formerly LMArena)reportedExpert promptsArena rating 1,3941,314
Arena (formerly LMArena)reportedInstruction followingArena rating 1,3591,281
Arena (formerly LMArena)reportedOverallArena rating 1,3691,318
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,3411,275
Artificial AnalysisreportedArtificial Analysis Coding Indexindex score 17.420.7
Artificial AnalysisreportedArtificial Analysis Intelligence Indexindex score 11.210.0
Artificial AnalysisreportedGPQA Diamond% of questions 75.9%68.8%
Artificial AnalysisreportedHumanity's Last Exam% of questions 12.6%11.0%
Artificial AnalysisreportedIFBench% of instructions 60.7%65.1%
Artificial AnalysisreportedLong-context reasoning (AA-LCR)% of questions 63.7%34.7%
Artificial AnalysisreportedTerminal-Bench 2.1% of tasks 6.7%13.9%
Artificial AnalysisreportedTerminal-Bench Hard% of tasks 9.8%10.6%
Artificial AnalysisreportedΤ-bench banking% of tasks 6.2%7.0%
Artificial AnalysisreportedΤ²-bench telecom% of tasks 41.5%60.2%
UGI LeaderboardreportedRequested-length error% off the requested word count 67.0%98.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.280.31
UGI LeaderboardreportedWriting scorescore out of 100 27.224.6

Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.