Inference
EN

Qwen3.8-27B vs Qwen3.8-Flash-Next

Change models
Qwen3.8-27BAlibaba
Qwen3.8-Flash-NextAlibaba
IndexAll categories, out of 100.
59.2 ±2.1
64.3 ±3.2 best
GapIndex points below the top.
−5.1
Top
CoverageIndex weight that has results.
100% best
95%
SpeedOutput tokens per second.
113/s best
55/s
EloLMArena rating and rank.
1436
—
ReleasedMonth of release.
2026-08
2026-08
Input / 1MUS dollars per 1M input tokens.
$0.15
—
Output / 1MUS dollars per 1M output tokens.
$1.88
—
1K chats2K tokens in, 500 out, each.
$1.24
—
1K documents20K tokens in, 1K out, each.
$4.88
—
1K agent tasks200K tokens in, 20K out, each.
$67.50
—
ContextMaximum tokens in one request.
1.05M best
262K
AgenticMulti-step tasks with tools.
57.4
68.5 best
CodingCode writing and repair.
60.3
62.1 best
ReasoningLogic problems and puzzles.
61.2
62.2 best
MultimodalTasks with images and text.
63.6
65.0 best
KnowledgeFacts and expert knowledge.
51.2
55.2 best
MultilingualTasks in many languages.
60.6
—
InstructionTasks with strict prompt rules.
68.5
73.8 best
MathMath problems.
56.3
59.4 best

Benchmarks

46 shared
Only Qwen3.8-27B: 54Only Qwen3.8-Flash-Next: 3
Qwen3.8-27B
Qwen3.8-Flash-Next
Artificial Analysis Humanity's Last ExamKnowledge
33.9%67 / 222
38.0% best55 / 222
Artificial Analysis Long Context ReasoningReasoning
82.0% best29 / 217
79.7%45 / 217
Critical Physics TasksReasoning
5.4%81 / 214
11.1% best61 / 214
Artificial Analysis GPQA DiamondKnowledge
90.5%32 / 199
92.3% best19 / 199
Artificial Analysis Omniscience AccuracyKnowledge
17.3%143 / 164
23.5% best112 / 164
Graduate-Level Google-Proof Q&AKnowledge
89.2%34 / 141
91.7% best21 / 141
GDPval-AA normalizedAgentic
45.9%38 / 133
56.4% best13 / 133
Artificial Analysis SciCodeCoding
46.6%67 / 114
50.6% best49 / 114
Artificial Analysis MMMU-ProMultimodal
76.3%46 / 113
79.8% best24 / 113
GPQA Diamond · BenchLMKnowledge
89.2%37 / 84
91.7% best21 / 84
SWE-bench ProCoding
61.7%27 / 83
62.5% best24 / 83
LiveBench Data AnalysisReasoning
76.6% best31 / 80
74.2%43 / 80
LiveBench ReasoningReasoning
80.0%58 / 80
87.4% best28 / 80
Humanity's Last ExamKnowledge
30.8%45 / 77
35.9% best39 / 77
LiveBench Agentic CodingAgentic
61.4%13 / 66
61.6% best12 / 66
LiveBench CodingCoding
75.7% best47 / 66
72.6%53 / 66
LiveBench Instruction FollowingInstruction
72.7%18 / 66
77.1% best5 / 66
LiveBench LanguageKnowledge
74.3%50 / 66
74.6% best49 / 66
LiveBench MathematicsMath
86.2% best49 / 66
85.8%50 / 66
DeepSWEAgentic
42.2%57 / 59
58.7% best44 / 59