Inference
EN

DeepSeek V4 Pro 0423 vs Kimi K2.7 Code

Change models
DeepSeek V4 Pro 0423DeepSeek
Kimi K2.7 CodeMoonshot AI
IndexAll categories, out of 100.
57.6 ±4.8
57.7 ±4.5 best
GapIndex points below the top.
−0.1 · no clear gap
Top
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
89/s best
35/s
EloLMArena rating and rank.
1451
—
ReleasedMonth of release.
2026-04
2026-06
Input / 1MUS dollars per 1M input tokens.
$0.892
$0.60 best
Output / 1MUS dollars per 1M output tokens.
$1.78 best
$3.50
1K chats2K tokens in, 500 out, each.
$2.68 best
$2.95
1K documents20K tokens in, 1K out, each.
$19.63
$15.50 best
1K agent tasks200K tokens in, 20K out, each.
$214.15
$190.00 best
ContextMaximum tokens in one request.
1.05M best
262K
AgenticMulti-step tasks with tools.
55.9
58.4 best
CodingCode writing and repair.
59.2
60.4 best
ReasoningLogic problems and puzzles.
54.6
54.6
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
59.9 best
58.8
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
52.7 best
44.2
MathMath problems.
57.6 best
57.0

Benchmarks

20 shared
Only DeepSeek V4 Pro 0423: 21Only Kimi K2.7 Code: 17
DeepSeek V4 Pro 0423
Kimi K2.7 Code
GPQA diamond · Epoch AIKnowledge
89.6% best44 / 217
87.9%59 / 217
OTIS Mock AIME 2024-2025Math
96.7% best36 / 200
95.6%40 / 200
LiveCodeBenchCoding
87.5% best13 / 142
82.0%57 / 142
Vibe Code Bench v1.1Coding
49.9% best66 / 123
47.2%73 / 123
SimpleQA VerifiedKnowledge
47.0% best37 / 91
36.5%57 / 91
Chess PuzzlesReasoning
19.0% best68 / 90
14.0%74 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
45.3%58 / 90
54.0% best52 / 90
SWE-benchCoding
77.4%35 / 87
78.2% best30 / 87
LiveBench Data AnalysisReasoning
74.5% best41 / 80
62.7%71 / 80
LiveBench ReasoningReasoning
82.7%51 / 80
82.8% best49 / 80
Code MigrationCoding
26.2% best46 / 77
25.4%48 / 77
Terminal-Bench 2.1 · Vals AIAgentic
50.2%58 / 75
67.0% best36 / 75
FrontierMath-Tier-4-v2-PrivateMath
2.4%65 / 72
12.2% best58 / 72
LiveBench Agentic CodingAgentic
42.6%57 / 66
45.7% best52 / 66
LiveBench CodingCoding
70.0%59 / 66
74.0% best51 / 66
LiveBench Instruction FollowingInstruction
62.4% best52 / 66
56.3%61 / 66
LiveBench LanguageKnowledge
78.1% best37 / 66
77.9%39 / 66
LiveBench MathematicsMath
90.7% best31 / 66
79.6%59 / 66
ProgramBenchCoding
0.0%34 / 65
0.0%34 / 65
SkillsBenchCoding
51.3% best21 / 35
50.0%23 / 35