Inference
EN

Qwen3.6-27B vs Qwen 3.6 Max (preview)

Change models
Qwen3.6-27BAlibaba
Qwen 3.6 Max (preview)Alibaba
IndexAll categories, out of 100.
52.1 ±2.4
57.2 ±5.0 best
GapIndex points below the top.
−5.1 · no clear gap
Top
CoverageIndex weight that has results.
100% best
85%
SpeedOutput tokens per second.
60/s
75/s best
EloLMArena rating and rank.
—
1446
ReleasedMonth of release.
2026-04
2026-04
Input / 1MUS dollars per 1M input tokens.
$0.32 best
$1.03
Output / 1MUS dollars per 1M output tokens.
$3.20 best
$6.16
1K chats2K tokens in, 500 out, each.
$2.24 best
$5.14
1K documents20K tokens in, 1K out, each.
$9.60 best
$26.70
1K agent tasks200K tokens in, 20K out, each.
$128.00 best
$328.64
ContextMaximum tokens in one request.
262K
262K
AgenticMulti-step tasks with tools.
56.7
59.2 best
CodingCode writing and repair.
52.8
57.0 best
ReasoningLogic problems and puzzles.
49.0
53.4 best
MultimodalTasks with images and text.
53.4
—
KnowledgeFacts and expert knowledge.
48.2
59.3 best
MultilingualTasks in many languages.
60.9
—
InstructionTasks with strict prompt rules.
43.2
60.6 best
MathMath problems.
51.0
53.2 best

Benchmarks

19 shared
Only Qwen3.6-27B: 45Only Qwen 3.6 Max (preview): 9
Qwen3.6-27B
Qwen 3.6 Max (preview)
Artificial Analysis Humanity's Last ExamKnowledge
23.1%104 / 222
30.8% best73 / 222
Artificial Analysis Long Context ReasoningReasoning
77.3%77 / 217
80.7% best37 / 217
GPQA diamond · Epoch AIKnowledge
84.8%72 / 217
87.4% best63 / 217
Critical Physics TasksReasoning
1.1%118 / 214
3.7% best93 / 214
OTIS Mock AIME 2024-2025Math
66.7%106 / 200
91.1% best58 / 200
Artificial Analysis GPQA DiamondKnowledge
84.2%83 / 199
88.8% best49 / 199
Artificial Analysis Omniscience AccuracyKnowledge
19.6%134 / 164
37.9% best60 / 164
τ²-Bench Tool-Agent-User EvaluationAgentic
94.2%26 / 149
95.9% best13 / 149
Artificial Analysis IFBenchInstruction
67.6%53 / 146
76.6% best11 / 146
CorpFin v2Finance
62.3%47 / 131
66.5% best18 / 131
SWE-benchCoding
70.0%61 / 87
72.8% best54 / 87
SWE-bench ProCoding
53.5%60 / 83
57.3% best43 / 83
Mystery Game PuzzlesReasoning
7.0%69 / 80
19.0% best46 / 80
Terminal-Bench 2.0 · Vals AIAgentic
44.9%24 / 66
51.7% best18 / 66
CaseLaw v2Legal
53.2% best40 / 53
47.9%50 / 53
Terminal-Bench 2.0 · BenchLMAgentic
59.3%18 / 41
65.4% best12 / 41
NL2RepoCoding
36.2%29 / 36
42.9% best22 / 36
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate DisciplinesKnowledge
66.0%14 / 22
73.9% best3 / 22
QwenClawBenchAgentic
53.4%7 / 10
59.0% best3 / 10