deepseek
DeepSeek: DeepSeek V4 Flash
DeepSeek V4 Flash, from DeepSeek, offers a context length of 1 million tokens and supports text input only. It is capable of handling reasoning tasks and integrating tools, though it does not provide structured output. With a benchmark blended score of 75.1 across four independent benchmarks, it performs well in various AI tasks but excels more in coding-related tasks. This model might be suitable for developers or professionals needing robust text-based reasoning abilities and tool integration at a relatively lower price point. Given its cost of $0.098 per million input tokens and $0.196 per million output tokens, it stands out as a cost-effective option compared to some more expensive alternatives, making it a strong contender for those prioritizing value in their AI model selection.
Quality Score
99/100
price + capability + benchmarks
Input Price
$0.10
per 1M tokens
Output Price
$0.20
per 1M tokens
Context Window
1,048,576
tokens
Benchmark results
Independent, published benchmarks. Blended score 75.1 across 4 benchmarks, last refreshed 2026-07-23. How scoring works →
| Benchmark | Measures | Score |
|---|---|---|
| AI Index | broad capability composite | 66.5 |
| AI Index Coding | software engineering tasks | 92.7 |
| AI Index Agentic | multi-step tool-using tasks | 51.3 |
| EQ-Bench | emotional understanding in dialogue | 72.5 |
- Model ID
- deepseek/deepseek-v4-flash
- Vendor
- deepseek
- Released
- April 2026
- Tokenizer
- DeepSeek
- Input Modalities
- text
- Output Modalities
- text
- Max Output
- default
- Tool Calling
- ✓ supported
- Structured Output
- ✓ supported
- Reasoning Mode
- ✓ supported
- Vision
- text only
- Audio
- no
- Moderated
- no
What it costs in practice
Computed from the current $0.10/M input and $0.20/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | under $0.01 |
| Classify 1,000 customer emails | 500k in / 50k out | $0.06 |
| A month of a busy support chatbot | 5M in / 2M out | $0.88 |
Price & spec history
Tracked daily by PicksByModel since 2026-07-17.
| Date | Input /M | Output /M | Context |
|---|---|---|---|
| 2026-07-23 | $0.10 | $0.20 | 1,048,576 |
| 2026-07-21 | $0.09 | $0.19 | 1,048,576 |
| 2026-07-17 | $0.10 | $0.20 | 1,048,576 |
Strong choice for
Code
Code Completion
Inline IDE-style autocomplete that has to feel instant.
Code
Code Documentation
Writing clear docstrings and READMEs that match the code.
Data
Bulk Data Labeling
Cheaply tagging thousands of items with consistent labels.
Writing
Short-Form Summarization
TL;DRs of articles and emails at scale.
Writing
Email Drafting
Cold emails, replies, and outreach at the right tone.
Writing
Transcript Cleanup
Turning raw ASR output into readable prose.
Category rankings
Where DeepSeek: DeepSeek V4 Flash places across the 68 categories it ranks in. How we rank →
| # | Category | Score |
|---|---|---|
| #1 | Code CompletionCode · of 25 ranked | 135 |
| #1 | Short-Form SummarizationWriting · of 25 ranked | 132 |
| #1 | Email DraftingWriting · of 25 ranked | 128 |
| #1 | Transcript CleanupWriting · of 25 ranked | 146 |
| #1 | Language LearningEducation · of 25 ranked | 128 |
| #1 | Chat CompanionPersonal · of 25 ranked | 132 |
| #2 | Bulk Data LabelingData · of 25 ranked | 135 |
| #2 | Customer SupportBusiness · of 25 ranked | 133 |
| #2 | Sales / Cold EmailBusiness · of 25 ranked | 124 |
| #2 | Job Application DraftingBusiness · of 25 ranked | 124 |
| #2 | Journaling HelperPersonal · of 25 ranked | 124 |
| #2 | Recipe GenerationPersonal · of 25 ranked | 124 |
| #2 | Dataset AnnotationResearch · of 25 ranked | 144 |
| #3 | Trivia & General KnowledgePersonal · of 25 ranked | 122 |
| #4 | Code DocumentationCode · of 25 ranked | 147 |
| #5 | Meeting NotesBusiness · of 25 ranked | 156 |
| #5 | Resume WritingBusiness · of 25 ranked | 118 |
| #6 | Regex WritingCode · of 25 ranked | 136 |
| #6 | CSV / Spreadsheet CleanupData · of 25 ranked | 156 |
| #6 | JSON ExtractionData · of 25 ranked | 147 |
| #6 | Long-Document SummarizationWriting · of 25 ranked | 164 |
| #6 | Blog Post WritingWriting · of 25 ranked | 141 |
| #6 | Language TranslationWriting · of 25 ranked | 136 |
| #6 | Character RoleplayPersonal · of 25 ranked | 153 |
| #6 | Trip PlanningPersonal · of 25 ranked | 141 |
| #6 | Tutoring for KidsPersonal · of 25 ranked | 136 |
| #8 | Code RefactoringCode · of 25 ranked | 167 |
| #8 | ETL ScriptingData · of 25 ranked | 156 |
| #8 | Medical Note SummarizationProfessional · of 25 ranked | 156 |
| #8 | History TutoringEducation · of 25 ranked | 147 |
| #8 | RAG PipelinesAgents · of 25 ranked | 167 |
| #8 | Long-Context Q&AAgents · of 25 ranked | 158 |
| #9 | Creative WritingWriting · of 25 ranked | 159 |
| #9 | Screenwriting & DialogWriting · of 25 ranked | 159 |
| #9 | Academic WritingWriting · of 25 ranked | 159 |
| #9 | Legal DraftingProfessional · of 25 ranked | 159 |
| #9 | Legal ResearchProfessional · of 25 ranked | 170 |
| #9 | Contract ReviewProfessional · of 25 ranked | 180 |
| #9 | Scientific ResearchProfessional · of 25 ranked | 170 |
| #9 | RFP ResponseBusiness · of 25 ranked | 177 |
| #9 | OKRs & Strategic PlanningBusiness · of 25 ranked | 159 |
| #9 | Fiction CollaboratorPersonal · of 25 ranked | 170 |
| #9 | Math ProofsResearch · of 25 ranked | 159 |
| #9 | Literature ReviewResearch · of 25 ranked | 170 |
| #10 | Code ReviewCode · of 25 ranked | 168 |
| #10 | Essay GradingEducation · of 25 ranked | 168 |
| #11 | Marketing CopyWriting · of 25 ranked | 130 |
| #11 | Standardized Test PrepEducation · of 25 ranked | 130 |
| #12 | SQL GenerationCode · of 25 ranked | 169 |
| #12 | CI/CD PipelinesCode · of 25 ranked | 144 |
| #12 | SOP / Process DocsBusiness · of 25 ranked | 136 |
| #13 | Unit Test GenerationCode · of 25 ranked | 154 |
| #13 | Physics TutoringEducation · of 25 ranked | 147 |
| #14 | Bug FixingCode · of 25 ranked | 179 |
| #14 | Scientific CodingResearch · of 25 ranked | 179 |
| #15 | Math TutoringEducation · of 25 ranked | 145 |
| #15 | Agent WorkflowsAgents · of 25 ranked | 189 |
| #15 | Coding AgentsAgents · of 25 ranked | 189 |
| #15 | Fitness CoachingPersonal · of 25 ranked | 133 |
| #15 | Cheap Bulk InferenceCost · of 25 ranked | 137 |
| #16 | Data AnalysisData · of 25 ranked | 165 |
| #16 | Financial AnalysisProfessional · of 25 ranked | 155 |
| #16 | Experiment DesignResearch · of 25 ranked | 154 |
| #16 | Self-Hosted / LocalCost · of 25 ranked | 117 |
| #18 | Function / Tool CallingAgents · of 25 ranked | 147 |
| #20 | Social Media PostsWriting · of 25 ranked | 119 |
| #20 | Voice Assistant BackendVoice · of 25 ranked | 123 |
| #22 | Real-Time ChatLatency · of 25 ranked | 117 |
Similar models
deepseek
DeepSeek: DeepSeek V4 Pro
$0.43 in / $0.87 out
1,048,576 ctx
99
deepseek
DeepSeek: DeepSeek V3.2
$0.27 in / $0.40 out
163,840 ctx
95
deepseek
DeepSeek: DeepSeek V3.2 Exp
$0.27 in / $0.41 out
163,840 ctx
95
deepseek
DeepSeek: DeepSeek V3.1
$0.25 in / $0.95 out
163,840 ctx
94
deepseek
DeepSeek: DeepSeek V3.1 Terminus
$0.27 in / $1.00 out
163,840 ctx
94
deepseek
DeepSeek: R1 0528
$0.50 in / $2.15 out
163,840 ctx
94
Quick answers
- How much does DeepSeek: DeepSeek V4 Flash cost?
- $0.10 per million input tokens and $0.20 per million output tokens.
- What is DeepSeek: DeepSeek V4 Flash's context window?
- 1,048,576 tokens, roughly 1,572 pages of text in a single request.
- Does DeepSeek: DeepSeek V4 Flash support tool calling?
- It supports tool calling, structured output, a reasoning mode.
- Can DeepSeek: DeepSeek V4 Flash process images?
- No, it is text-only on the input side.