Claude 3.7 Sonnet vs Qwen3 235B-A22B

Change models
IndexAll categories, out of 100.
37.6 ±4.0
37.8 ±8.8 best
GapIndex points below the top.
−0.2 · no clear gap
Top
CoverageIndex weight that has results.
95% best
55%
SpeedOutput tokens per second.
—
44/s
EloLMArena rating and rank.
1299
1366 best
ReleasedMonth of release.
2025-02
2025-04
Input / 1MUS dollars per 1M input tokens.
—
$0.455
Output / 1MUS dollars per 1M output tokens.
—
$1.82
1K chats2K tokens in, 500 out, each.
—
$1.82
1K documents20K tokens in, 1K out, each.
—
$10.92
1K agent tasks200K tokens in, 20K out, each.
—
$127.40
ContextMaximum tokens in one request.
200K best
131K
AgenticMulti-step tasks with tools.
41.4
—
CodingCode writing and repair.
41.8
43.7 best
ReasoningLogic problems and puzzles.
39.6
—
MultimodalTasks with images and text.
39.8
—
KnowledgeFacts and expert knowledge.
40.1
40.9 best
MultilingualTasks in many languages.
—
41.8
InstructionTasks with strict prompt rules.
38.0 best
30.9
MathMath problems.
29.6
31.8 best

Benchmarks

16 shared
Only Claude 3.7 Sonnet: 26Only Qwen3 235B-A22B: 8
Claude 3.7 Sonnet
Qwen3 235B-A22B
76.8% best102 / 220
70.7%116 / 220
80.0%91 / 149
80.2% best89 / 149
68.0% best97 / 144
62.9%108 / 144
TaxEval v2Finance
72.4% best62 / 143
70.6%89 / 143
56.7%111 / 142
70.6% best86 / 142
67.4%100 / 135
70.2% best92 / 135
MMLU ProKnowledge
80.7%85 / 135
81.2% best82 / 135
AIMEMath
22.3%79 / 95
84.0% best40 / 95
88.1% best18 / 88
68.9%34 / 88
MGSMMultilingual
92.4%18 / 74
92.5% best17 / 74
90.8% best25 / 69
83.2%45 / 69
0.0%33 / 64
0.0%33 / 64
76.8%43 / 59
94.6% best8 / 59
82.2%27 / 29
91.2% best22 / 29
23.3%27 / 28
40.1% best24 / 28
MMMLUKnowledge
83.2% best10 / 15
79.8%13 / 15