Claude Opus 4.7 vs Claude Sonnet 4.6

Change models
IndexAll categories, out of 100.
62.6 ±2.7 best
56.2 ±2.5
GapIndex points below the top.
Top
−6.4
CoverageIndex weight that has results.
95%
100% best
SpeedOutput tokens per second.
46/s best
38/s
EloLMArena rating and rank.
1483 best
1458
ReleasedMonth of release.
2026-04
2026-02
Input / 1MUS dollars per 1M input tokens.
$5.00
$3.00 best
Output / 1MUS dollars per 1M output tokens.
$25.00
$15.00 best
Batch output / 1MBatch price per 1M output tokens.
$12.50
$7.50 best
1K chats2K tokens in, 500 out, each.
$22.50
$13.50 best
1K documents20K tokens in, 1K out, each.
$125.00
$75.00 best
1K agent tasks200K tokens in, 20K out, each.
$1,500.00
$900.00 best
ContextMaximum tokens in one request.
1M
1M
AgenticMulti-step tasks with tools.
61.3 best
58.6
CodingCode writing and repair.
63.8 best
58.2
ReasoningLogic problems and puzzles.
61.9 best
55.6
MultimodalTasks with images and text.
60.8 best
53.3
KnowledgeFacts and expert knowledge.
64.5 best
56.7
MultilingualTasks in many languages.
—
56.5
InstructionTasks with strict prompt rules.
52.2 best
28.0
MathMath problems.
63.2 best
53.4

Benchmarks

54 shared
Only Claude Opus 4.7: 31Only Claude Sonnet 4.6: 27
Claude Opus 4.7
Claude Sonnet 4.6
86.4% best70 / 220
78.8%96 / 220
42.3% best39 / 216
13.3%126 / 216
78.7% best61 / 211
68.3%117 / 211
12.0% best56 / 207
0.9%119 / 207
86.7% best72 / 203
71.1%101 / 203
91.4% best26 / 193
79.9%95 / 193
48.9% best27 / 170
38.6%61 / 170
85.3% best22 / 149
82.1%75 / 149
94.2% best7 / 144
89.9%33 / 144
TaxEval v2Finance
75.3%19 / 143
77.1% best4 / 143
88.6% best44 / 143
79.5%67 / 143
85.1% best35 / 142
82.1%56 / 142
58.6% best61 / 140
41.2%95 / 140
90.2% best25 / 135
85.6%45 / 135
MMLU ProKnowledge
89.9% best8 / 135
87.3%29 / 135
CorpFin v2Finance
66.1% best21 / 131
65.3%28 / 131
71.0% best54 / 130
51.5%72 / 130
87.6% best6 / 114
79.6%23 / 114
78.8% best32 / 109
70.6%70 / 109
38.5%44 / 98
38.5%44 / 98