DeepSeek model product
DeepSeek-V4-Flash
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version with substantially stronger agentic capabilities.
Updated Aug 12, 2026. Default version: DeepSeek-V4-Flash-0731
LLMBoard score78.3DeepSeek-V4-Flash-0731
Coverage20%7 benchmark families
Context window1MTokens
Official input price$0.14DeepSeek API
DeepSeek-V4-Flash Specifications
Technical details for the model's default version.
- Version
- DeepSeek-V4-Flash-0731
- Released
- Jul 31, 2026
- Knowledge cutoff
- May 1, 2025
- Parameters
- 304B
- Context window
- 1M
- Max output
- 384K
- Inputs
- text
- Outputs
- text
- Open weights
- Yes
- License
- MIT
DeepSeek-V4-Flash Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
DeepSeek-V4-Flash Benchmark Results
Benchmark scores for DeepSeek-V4-Flash-0731.
| DSBench-FullStack | 68.7% | 01 | 1 | 100.0% | C | |
| DSBench-Hard | 59.6% | 01 | 1 | 100.0% | C | |
| NL2Repo | 54.2% | 02 | 14 | 92.3% | C | |
| AutomationBench | 25.1% | 03 | 8 | 71.4% | C | |
| Toolathlon | 70.3% | 03 | 31 | 93.3% | C | |
| Agents' Last Exam | 25.2% | 05 | 5 | 0.0% | C | |
| CyberGym | 76.7% | 05 | 11 | 60.0% | C | |
| DeepSWE | 54.4% | 05 | 10 | 55.6% | C | |
| Terminal-Bench 2.1 | 82.7% | 09 | 19 | 55.6% | C |
DeepSeek-V4-Flash Arena Results
Preference and agent-evaluation results for the default version.
| webdev | overall | 08 | 1584.9 | 2,605 | N/A | |
| agent task outcome explicit | overall | 09 | 0.1 | N/A | 10.6K | |
| agent tool hallucination | overall | 15 | 0.0 | N/A | 1.4M | |
| agent | overall | 21 | 0.0 | N/A | 1.5M | |
| agent steerability | overall | 23 | 0.0 | N/A | 13.5K | |
| agent praise complaint | overall | 26 | 0.0 | N/A | 3.7K | |
| agent bash recovery steps | overall | 27 | 0.0 | N/A | 52.6K | |
| webdev | overall | 53 | 1431.3 | 9,493 | N/A | |
| text style control | overall | 79 | 1438.1 | 48,812 | N/A | |
| text | overall | 87 | 1423.8 | 48,812 | N/A | |
| text factuality | overall | 87 | 1433.7 | 48,725 | N/A |
DeepSeek-V4-Flash Runtime Performance
Provider-specific output speed and catalog latency for DeepSeek-V4-Flash-0731. Runtime does not affect the capability score.
| Novita | 10.973 tok/s | 4.091 s | 1M | 393.2K | |
| DeepInfra | 6.415 tok/s | 15.272 s | 1M | 65.5K | |
| Fireworks | 1.414 tok/s | 5.906 s | 1M | 65.5K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
DeepSeek-V4-Flash Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.14 input, $0.28 output per 1M
- Official provider
- DeepSeek
- Lowest third-party
- From $0.08 input, $0.252 output per 1M via DigitalOcean
- Tracked offerings
- 30
| Alibaba Token Plan | deepseek-v4-flash-0731 | global | N/A | N/A | 1M | |
| Umans AI Coding Plan | umans-deepseek-v4-flash-0731 | global | N/A | N/A | 1M | |
| Alibaba Token Plan (China) | deepseek-v4-flash-0731 | global | N/A | N/A | 1M | |
| DigitalOcean | deepseek-v4-flash-0731 | global | $0.08 | $0.252 | 1M | |
| OpenRouter | deepseek/deepseek-v4-flash-0731 | global | $0.08 | $0.18 | 1M | |
| Deep Infra | deepseek-ai/DeepSeek-V4-Flash-0731 | global | $0.08 | $0.18 | 1M | |
| CrofAI | deepseek-v4-flash-0731 | global | $0.12 | $0.21 | 1M | |
| Weights & Biases | deepseek-ai/DeepSeek-V4-Flash-0731 | global | $0.13 | $0.28 | 262.1K | |
| Baseten | deepseek-ai/DeepSeek-V4-Flash-0731 | global | $0.13 | $0.26 | 1M | |
| Fireworks AI | accounts/fireworks/models/deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| NanoGPT | deepseek/deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| OpenCode Zen | deepseek-v4-flash | global | $0.14 | $0.28 | 1M | |
| Ambient | deepseek/deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| EmpirioLabs AI | deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| Together AI | deepseek-ai/DeepSeek-V4-Flash-0731 | global | $0.14 | $0.28 | 1M | |
| Umans AI | umans-deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| DeepSeek | deepseek-v4-flash | global | $0.14 | $0.28 | 1M | |
| Requesty | deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| Kilo Gateway | deepseek/deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| Merge Gateway | deepseek/deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| Hugging Face | deepseek-ai/DeepSeek-V4-Flash-0731 | global | $0.14 | $0.28 | 1M | |
| Vivgrid | deepseek-v4-flash | global | $0.15 | $0.30 | 1M | |
| Charm Hyper | deepseek-v4-flash-0731 | global | $0.1524 | $0.3049 | 1M | |
| GreenPT | deepseek-v4-flash-0731 | global | $0.1596 | $0.399 | 1M | |
| Venice AI | deepseek-v4-flash-0731 | global | $0.175 | $0.35 | 1M | |
| Alibaba | deepseek-v4-flash-0731 | global | $0.20 | $0.40 | 1M | |
| Vercel AI Gateway | deepseek/deepseek-v4-flash-0731 | global | $0.20 | $0.40 | 1M | |
| TensorX | deepseek/deepseek-v4-flash-0731 | global | $0.25 | $0.30 | 1M | |
| Cortecs | deepseek-v4-flash-0731 | global | $0.25 | $0.30 | 1M |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
DeepSeek-V4-Flash Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| DeepSeek-V4-Flash-0731 | 78.3 | 304B | 1M | 384K | Yes | MIT | |
| DeepSeek V4 Flash | N/A | N/A | 1M | 384K | Yes | Unspecified | |
| DeepSeek-V4-Flash-Max | N/A | 284B | 1M | 393.2K | No | MIT |
What is DeepSeek-V4-Flash?
Key information about DeepSeek-V4-Flash and its available data.
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version with substantially stronger agentic capabilities. It retains the V4-Flash architecture and adds an attached DSpark speculative-decoding module for faster inference.
The open-weight model supports a 1M-token context window and low, high, and max reasoning-effort levels, and DeepSeek reports that it outperforms the much larger DeepSeek-V4-Pro Preview across its published agent benchmark suite.
Data as of 2026-08-11.
DeepSeek-V4-Flash vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to DeepSeek-V4-Flash
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
DE
DeepSeek-V4-Pro
DeepSeek
DE
DeepSeek-V3.2
DeepSeek
DE
DeepSeek-R1
DeepSeek
DE
DeepSeek-V3.1
DeepSeek
DE
DeepSeek-V3
DeepSeek
DE
DeepSeek-R1-Distill-Qwen
DeepSeek
FAQ
Common questions about DeepSeek-V4-Flash.
When was DeepSeek-V4-Flash released?
DeepSeek-V4-Flash's default version was released on Jul 31, 2026.
How much does DeepSeek-V4-Flash cost?
DeepSeek-V4-Flash's official API price is $0.14 per million input tokens and $0.28 per million output tokens via DeepSeek. The lowest tracked third-party offer starts at $0.08 input and $0.252 output via DigitalOcean.
Who created DeepSeek-V4-Flash?
DeepSeek-V4-Flash was created by DeepSeek.
What is the context window for DeepSeek-V4-Flash?
The default version has a 1M token context window.
Is DeepSeek-V4-Flash open weight?
Yes. The default version is marked as open weight under MIT.
How many API providers offer DeepSeek-V4-Flash?
30 provider offerings are linked to the default version.
What models should I compare DeepSeek-V4-Flash with?
Nearby ranked alternatives include Gemini 3.5 Flash, Qwen3.7 Max, Seed 2.1 Pro.