Compare Models
DeepSeek V4 Flash 0731

Release Date: Jul 31, 2026

Developer DeepSeek 🇨🇳
Context Window 1M
Max Output Tokens 384k
Token Costs (in/out) $0.14/0.28
Weights Open
Input Modalities

Accuracy

53.57 % ± 1.21

Cost / Test (Vals Index)

$ 0.219

Latency

34 min 55 s

Vals Index
BenchmarksAccuracyRankings
Vals Index

0.0%

±1.21
16/46
Code Migration

0.0%

±4.38
12/49
CyberBench

0.0%

±5.59
5/22
EMB

0.0%

±3.39
21/46
Finance Agent (v2)

0.0%

±0.47
21/49
Legal Research Bench

0.0%

±3.19
21/49
MedCode

0.0%

±2.15
38/85
MedScribe

0.0%

±1.97
38/84
ProofBench v1.1

0.0%

±4.96
10/23
TaxEval v2

0.0%

±0.90
85/140
Vibe Code Bench v1.1

0.0%

±3.44
12/84
Harvey's Legal Agent Benchmark

0.0%

±2.18
10/50
GPQA Diamond

0.0%

±1.69
25/133
LiveCodeBench

0.0%

±0.97
15/138
LegalBench

0.0%

±0.50
99/137
MMLU Pro

0.0%

±0.34
44/133
SkillsBench

0.0%

±4.43
19/28
SWE-bench

0.0%

±1.41
8/83
Terminal-Bench 2.1

0.0%

±1.35
22/54
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : DeepSeek
Temperature: 1
Top P: Default
Top K: Default
Max Output Tokens: 384,000
Reasoning Effort: high