Claude 3.7 Sonnet vs DeepSeek V3.1
Change modelsClaude 3.7 SonnetAnthropic
DeepSeek V3.1DeepSeek
IndexAll categories, out of 100.
37.7 ±4.0
37.9 ±5.8 best
GapIndex points below the top.
−0.2 · no clear gap
Top
CoverageIndex weight that has results.
95% best
85%
SpeedOutput tokens per second.
—
98/s
EloLMArena rating and rank.
1299
1420 best
ReleasedMonth of release.
2025-02
2025-08
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
200K best
128K
AgenticMulti-step tasks with tools.
40.7 best
38.0
CodingCode writing and repair.
42.1
43.0 best
ReasoningLogic problems and puzzles.
39.8 best
38.2
MultimodalTasks with images and text.
40.3
—
KnowledgeFacts and expert knowledge.
40.0
41.5 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
38.1 best
24.3
MathMath problems.
30.0 best
29.9
Benchmarks
10 sharedOnly Claude 3.7 Sonnet: 32Only DeepSeek V3.1: 7
Claude 3.7 Sonnet
DeepSeek V3.1