DeepSeek model product
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.
Updated Aug 10, 2026. Default version: DeepSeek-V4-Flash-0731
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version DeepSeek-V4-Flash-Max.
| CodeForces | 1.0 | 1 | 16 | 100.0% | C | |
| CorpusQA 1M | 0.6 | 2 | 2 | 0.0% | C | |
| LiveCodeBench | 0.9 | 2 | 73 | 98.6% | C | |
| MathArena Apex | 0.9 | 2 | 7 | 83.3% | C | |
| MRCR 1M | 0.8 | 2 | 3 | 50.0% | C | |
| HMMT Feb 26 | 0.9 | 3 | 11 | 80.0% | C | |
| CSimpleQA | 0.8 | 4 | 7 | 50.0% | C | |
| IMO-AnswerBench | 0.9 | 6 | 19 | 72.2% | C | |
| MMLU-Pro | 0.9 | 11 | 129 | 92.2% | C | |
| SWE-bench Multilingual | 0.7 | 12 | 34 | 66.7% | C | |
| SWE-Bench Verified | 0.8 | 16 | 104 | 85.4% | C | |
| MCP Atlas | 0.7 | 19 | 30 | 37.9% | C | |
| Toolathlon | 0.5 | 19 | 31 | 40.0% | C | |
| SimpleQA | 0.3 | 23 | 46 | 51.1% | C | |
| Humanity's Last Exam | 0.5 | 27 | 92 | 71.4% | C | |
| Terminal-Bench 2.0 | 0.6 | 27 | 49 | 45.8% | C | |
| BrowseComp | 0.7 | 28 | 58 | 52.6% | C | |
| GPQA | 0.9 | 30 | 233 | 87.5% | C | |
| SWE-Bench Pro | 0.5 | 38 | 44 | 13.9% | C |
Preference and agent-evaluation signals from published Arena datasets.
The default version has no published Arena rows, or its source alias has not been resolved.
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| Alibaba Token Plan | deepseek-v4-flash-0731 | global | N/A | N/A | 1M | |
| Alibaba Token Plan (China) | deepseek-v4-flash-0731 | global | N/A | N/A | 1M | |
| CrofAI | deepseek-v4-flash-0731 | global | $0.12 | $0.21 | 1M | |
| EmpirioLabs AI | deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| DigitalOcean | deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| Requesty | deepseek-v4-flash-0731 | global | $0.14 | $0.28 | 1M | |
| GreenPT | deepseek-v4-flash-0731 | global | $0.1596 | $0.399 | 1M | |
| Venice AI | deepseek-v4-flash-0731 | global | $0.175 | $0.35 | 1M | |
| Charm Hyper | deepseek-v4-flash-0731 | global | $0.20 | $0.40 | 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.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| DeepSeek-V4-Flash-0731 | N/A | 304B | 1M | 384K | Yes | MIT | |
| DeepSeek V4 Flash | N/A | N/A | 1M | 384K | Yes | Unspecified | |
| DeepSeek-V4-Flash-Max | 71.7 | 284B | 1M | 393.2K | No | MIT |
A concise description based on the published model registry.
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.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about DeepSeek-V4-Flash.
DeepSeek-V4-Flash's default version was released on Jul 31, 2026.
No official standard PAYG price is currently available for DeepSeek-V4-Flash. The lowest tracked third-party offer starts at $0.12 input and $0.21 output via CrofAI.
DeepSeek-V4-Flash is published under DeepSeek in the model registry.
The default version has a 1M token context window.
Yes. The default version is marked as open weight under MIT.
10 published provider offerings are linked to the default version.
Nearby ranked alternatives include Qwen3.6 Plus, Qwen3.6 27B, Nemotron 3 Ultra.