Inference
EN

Claude Sonnet 5 vs Qwen3.8-Flash-Next

Change models
Claude Sonnet 5Anthropic
Qwen3.8-Flash-NextAlibaba
IndexAll categories, out of 100.
64.2 ±2.7
64.3 ±3.2 best
GapIndex points below the top.
−0.1 · no clear gap
Top
CoverageIndex weight that has results.
95%
95%
SpeedOutput tokens per second.
75/s best
55/s
EloLMArena rating and rank.
1442
—
ReleasedMonth of release.
2026-06
2026-08
Input / 1MUS dollars per 1M input tokens.
$2.00
—
Output / 1MUS dollars per 1M output tokens.
$10.00
—
Batch output / 1MBatch price per 1M output tokens.
$5.00
—
1K chats2K tokens in, 500 out, each.
$9.00
—
1K documents20K tokens in, 1K out, each.
$50.00
—
1K agent tasks200K tokens in, 20K out, each.
$600.00
—
ContextMaximum tokens in one request.
1M best
262K
AgenticMulti-step tasks with tools.
65.3
68.5 best
CodingCode writing and repair.
65.3 best
62.1
ReasoningLogic problems and puzzles.
66.5 best
62.2
MultimodalTasks with images and text.
59.6
65.0 best
KnowledgeFacts and expert knowledge.
59.0 best
55.2
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
55.8
73.8 best
MathMath problems.
64.8 best
59.4

Benchmarks

30 shared
Only Claude Sonnet 5: 56Only Qwen3.8-Flash-Next: 19
Claude Sonnet 5
Qwen3.8-Flash-Next
Artificial Analysis Humanity's Last ExamKnowledge
41.3% best41 / 222
38.0%55 / 222
Artificial Analysis Long Context ReasoningReasoning
82.0% best29 / 217
79.7%45 / 217
Critical Physics TasksReasoning
16.9% best43 / 214
11.1%61 / 214
Artificial Analysis GPQA DiamondKnowledge
91.1%27 / 199
92.3% best19 / 199
Artificial Analysis Omniscience AccuracyKnowledge
40.1% best51 / 164
23.5%112 / 164
GDPval-AA normalizedAgentic
48.2%29 / 133
56.4% best13 / 133
Artificial Analysis SciCodeCoding
54.3% best34 / 114
50.6%49 / 114
Artificial Analysis MMMU-ProMultimodal
77.3%40 / 113
79.8% best24 / 113
SWE-bench ProCoding
63.2% best22 / 83
62.5%24 / 83
LiveBench Data AnalysisReasoning
71.7%51 / 80
74.2% best43 / 80
LiveBench ReasoningReasoning
88.7% best20 / 80
87.4%28 / 80
Humanity's Last ExamKnowledge
57.4% best10 / 77
35.9%39 / 77
LiveBench Agentic CodingAgentic
59.4%17 / 66
61.6% best12 / 66
LiveBench CodingCoding
80.7% best16 / 66
72.6%53 / 66
LiveBench Instruction FollowingInstruction
63.9%45 / 66
77.1% best5 / 66
LiveBench LanguageKnowledge
75.0% best47 / 66
74.6%49 / 66
LiveBench MathematicsMath
92.9% best22 / 66
85.8%50 / 66
Agent Arena steerabilityAgentic
2.8 best20 / 54
−2.636 / 54
Agent Arena task outcomeAgentic
−0.633 / 54
0.3 best31 / 54
Humanity's Last Exam without toolsKnowledge
43.2% best22 / 53
35.9%32 / 53