Claude Opus 4.5 Thinking vs Qwen 3.6 Max (preview)

Change models
IndexAll categories, out of 100.
55.7 ±4.2
56.4 ±5.0 best
GapIndex points below the top.
−0.7 · no clear gap
Top
CoverageIndex weight that has results.
95% best
85%
SpeedOutput tokens per second.
—
75/s
EloLMArena rating and rank.
—
1446
ReleasedMonth of release.
2025-11
2026-04
Input / 1MUS dollars per 1M input tokens.
—
$1.03
Output / 1MUS dollars per 1M output tokens.
—
$6.16
1K chats2K tokens in, 500 out, each.
—
$5.14
1K documents20K tokens in, 1K out, each.
—
$26.70
1K agent tasks200K tokens in, 20K out, each.
—
$328.64
ContextMaximum tokens in one request.
200K
262K best
AgenticMulti-step tasks with tools.
59.2 best
58.5
CodingCode writing and repair.
53.9
56.3 best
ReasoningLogic problems and puzzles.
53.6 best
52.5
MultimodalTasks with images and text.
56.7
—
KnowledgeFacts and expert knowledge.
58.5
58.7 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
43.0
59.6 best
MathMath problems.
56.5 best
52.7

Benchmarks

13 shared
Only Claude Opus 4.5 Thinking: 24Only Qwen 3.6 Max (preview): 15
Claude Opus 4.5 Thinking
Qwen 3.6 Max (preview)
30.1%77 / 216
30.8% best72 / 216
77.3%76 / 211
80.7% best37 / 211
4.6% best85 / 207
3.7%91 / 207
86.6%60 / 193
88.8% best48 / 193
46.6% best36 / 170
37.9%63 / 170
89.5%41 / 143
95.9% best13 / 143
58.0%62 / 140
76.6% best10 / 140
CorpFin v2Finance
65.1%32 / 131
66.5% best18 / 131
SWE-benchCoding
76.4% best37 / 87
72.8%54 / 87
53.9% best16 / 66
51.7%18 / 66
62.6% best18 / 53
47.9%50 / 53
58.8% best8 / 51
52.8%22 / 51
47.0% best11 / 22
43.9%17 / 22