DeepSeek model product
Open MoE flagship with million-token context for coding and long agent runs
Updated Aug 10, 2026. Default version: DeepSeek V4 Pro
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-Pro-Max.
| CorpusQA 1M | 0.6 | 1 | 2 | 100.0% | C | |
| CSimpleQA | 0.8 | 1 | 7 | 100.0% | C | |
| LiveCodeBench | 0.9 | 1 | 73 | 100.0% | C | |
| MathArena Apex | 0.9 | 1 | 7 | 100.0% | C | |
| MRCR 1M | 0.8 | 1 | 3 | 100.0% | C | |
| CodeForces | 1.0 | 2 | 16 | 93.3% | C | |
| HMMT Feb 26 | 1.0 | 2 | 11 | 90.0% | C | |
| IMO-AnswerBench | 0.9 | 5 | 19 | 77.8% | C | |
| MMLU-Pro | 0.9 | 6 | 129 | 96.1% | C | |
| SWE-bench Multilingual | 0.8 | 8 | 34 | 78.8% | C | |
| SWE-Bench Verified | 0.8 | 8 | 104 | 93.2% | C | |
| SimpleQA | 0.6 | 9 | 46 | 82.2% | C | |
| FrontierSWE | 0.3 | 12 | 15 | 21.4% | B | |
| Toolathlon | 0.5 | 12 | 31 | 63.3% | C | |
| FrontierCode 1.1 | 0.2 | 13 | 15 | 14.3% | B | |
| Terminal-Bench 2.0 | 0.7 | 14 | 49 | 72.9% | C | |
| MCP Atlas | 0.7 | 15 | 30 | 51.7% | C | |
| BrowseComp | 0.8 | 17 | 58 | 71.9% | C | |
| LiveBench | 0.7 | 20 | 38 | 48.6% | B | |
| Humanity's Last Exam | 0.5 | 23 | 92 | 75.8% | C | |
| GPQA | 0.9 | 24 | 233 | 90.1% | C | |
| SWE-Bench Pro | 0.6 | 30 | 44 | 32.6% | C |
Preference and agent-evaluation signals from published Arena datasets.
| text factuality | japanese | 10 | 1460.2 | 441 | N/A | |
| text factuality | korean | 10 | 1423.0 | 483 | N/A | |
| text factuality | german | 11 | 1440.8 | 397 | N/A | |
| text factuality | spanish | 19 | 1454.6 | 1,194 | N/A | |
| text style control | japanese | 20 | 1461.9 | 608 | N/A | |
| text | japanese | 22 | 1449.8 | 608 | N/A | |
| text style control | german | 23 | 1471.4 | 882 | N/A | |
| agent steerability | overall | 24 | 0.0 | N/A | 32.4K | |
| agent | overall | 26 | -0.0 | N/A | 1.3M | |
| agent bash recovery steps | overall | 26 | 0.0 | N/A | 44.6K | |
| text | german | 27 | 1460.4 | 882 | N/A | |
| text | polish | 27 | 1459.8 | 1,007 | N/A | |
| text factuality | french | 27 | 1483.4 | 1,538 | N/A | |
| text factuality | industry medicine and healthcare | 27 | 1489.6 | 2,836 | N/A | |
| text | creative writing | 28 | 1445.4 | 8,475 | N/A | |
| agent praise complaint | overall | 29 | -0.0 | N/A | 7.6K | |
| text factuality | creative writing | 29 | 1446.0 | 7,471 | N/A | |
| text factuality | multi turn | 29 | 1476.0 | 8,019 | N/A | |
| text | multi turn | 30 | 1466.5 | 9,251 | N/A | |
| agent tool hallucination | overall | 32 | 0.0 | N/A | 1.2M | |
| agent task outcome explicit | overall | 33 | -0.0 | N/A | 19.6K | |
| text | industry medicine and healthcare | 34 | 1464.1 | 3,521 | N/A | |
| text | industry writing and literature and language | 34 | 1447.1 | 12,369 | N/A | |
| text factuality | industry entertainment and sports and media | 34 | 1437.3 | 10,478 | N/A | |
| text factuality | industry life and physical and social science | 34 | 1482.1 | 7,134 | N/A | |
| text factuality | industry writing and literature and language | 34 | 1449.3 | 11,754 | N/A | |
| text | english | 35 | 1457.5 | 22,730 | N/A | |
| text | industry entertainment and sports and media | 35 | 1430.7 | 11,151 | N/A | |
| text | industry life and physical and social science | 35 | 1468.3 | 8,098 | N/A | |
| text factuality | industry legal and government | 35 | 1475.4 | 3,264 | N/A | |
| text | industry legal and government | 36 | 1460.6 | 3,980 | N/A | |
| text | spanish | 36 | 1455.3 | 1,553 | N/A | |
| text | industry business and management and financial operations | 37 | 1443.5 | 10,136 | N/A | |
| text factuality | chinese | 37 | 1497.4 | 2,179 | N/A | |
| text factuality | longer query | 37 | 1473.5 | 21,836 | N/A | |
| text style control | industry writing and literature and language | 37 | 1452.2 | 12,369 | N/A | |
| text | exclude ties | 38 | 1448.1 | 38,848 | N/A | |
| text | french | 38 | 1468.3 | 1,999 | N/A | |
| text | overall | 39 | 1450.0 | 51,266 | N/A | |
| text | hard prompts english | 39 | 1463.8 | 15,953 | N/A | |
| text | instruction following | 39 | 1444.3 | 17,416 | N/A | |
| text | korean | 39 | 1412.1 | 981 | N/A | |
| text factuality | expert | 39 | 1485.9 | 4,140 | N/A | |
| text factuality | instruction following | 39 | 1455.0 | 16,836 | N/A | |
| text style control | creative writing | 39 | 1446.0 | 8,475 | N/A | |
| text style control | polish | 39 | 1464.1 | 1,007 | N/A | |
| text | longer query | 40 | 1457.5 | 22,336 | N/A | |
| text factuality | industry business and management and financial operations | 40 | 1458.2 | 9,128 | N/A | |
| text factuality | industry mathematical | 40 | 1445.8 | 2,222 | N/A | |
| text style control | french | 40 | 1481.4 | 1,999 | N/A | |
| text style control | multi turn | 40 | 1474.4 | 9,251 | N/A | |
| text | russian | 41 | 1449.3 | 5,394 | N/A | |
| text factuality | coding | 41 | 1504.8 | 14,678 | N/A | |
| text style control | korean | 41 | 1421.3 | 981 | N/A | |
| text | hard prompts | 42 | 1460.8 | 33,976 | N/A | |
| text factuality | hard prompts | 42 | 1477.3 | 33,934 | N/A | |
| text factuality | hard prompts english | 42 | 1487.3 | 15,390 | N/A | |
| text style control | industry legal and government | 42 | 1470.4 | 3,980 | N/A | |
| webdev | webdev-react | 42 | 1440.6 | 8,296 | N/A | |
| text | non english | 43 | 1439.0 | 28,536 | N/A | |
| text factuality | overall | 43 | 1453.9 | 51,202 | N/A | |
| text factuality | exclude ties | 43 | 1456.4 | 38,797 | N/A | |
| text factuality | non english | 43 | 1441.2 | 28,503 | N/A | |
| text style control | spanish | 43 | 1455.8 | 1,553 | N/A | |
| webdev | overall | 43 | 1446.3 | 11,908 | N/A | |
| webdev | webdev | 43 | 1446.3 | 11,908 | N/A | |
| text style control | instruction following | 44 | 1452.6 | 17,416 | N/A | |
| text style control | longer query | 44 | 1472.3 | 22,336 | N/A | |
| text factuality | industry software and it services | 45 | 1487.8 | 20,773 | N/A | |
| text style control | industry entertainment and sports and media | 45 | 1435.3 | 11,151 | N/A | |
| text style control | industry life and physical and social science | 45 | 1480.0 | 8,098 | N/A | |
| text | industry software and it services | 46 | 1468.4 | 21,041 | N/A | |
| text factuality | english | 46 | 1463.3 | 22,689 | N/A | |
| text factuality | math | 46 | 1441.2 | 2,147 | N/A | |
| text style control | industry medicine and healthcare | 47 | 1476.6 | 3,521 | N/A | |
| text style control | industry business and management and financial operations | 48 | 1459.8 | 10,136 | N/A | |
| text | expert | 49 | 1465.0 | 4,950 | N/A | |
| text style control | hard prompts | 49 | 1479.8 | 33,976 | N/A | |
| text style control | hard prompts english | 49 | 1483.2 | 15,953 | N/A | |
| webdev | webdev-html | 49 | 1432.5 | 1,263 | N/A | |
| text | coding | 50 | 1469.9 | 15,292 | N/A | |
| text style control | overall | 50 | 1457.1 | 51,266 | N/A | |
| text style control | exclude ties | 50 | 1460.4 | 38,848 | N/A | |
| text style control | russian | 50 | 1456.8 | 5,394 | N/A | |
| text | chinese | 51 | 1486.8 | 2,670 | N/A | |
| text | industry mathematical | 51 | 1453.8 | 2,813 | N/A | |
| text factuality | russian | 51 | 1454.5 | 4,439 | N/A | |
| text style control | english | 51 | 1465.9 | 22,730 | N/A | |
| text style control | non english | 51 | 1445.2 | 28,536 | N/A | |
| text style control | expert | 52 | 1481.8 | 4,950 | N/A | |
| text style control | industry mathematical | 52 | 1459.9 | 2,813 | N/A | |
| text style control | industry software and it services | 52 | 1491.4 | 21,041 | N/A | |
| text style control | chinese | 53 | 1495.6 | 2,670 | N/A | |
| text style control | coding | 56 | 1501.0 | 15,292 | N/A | |
| text | math | 59 | 1441.4 | 2,732 | N/A | |
| text style control | math | 63 | 1443.8 | 2,732 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
The default version has no provider offering with current input or output token prices.
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 Pro | N/A | N/A | 1M | 384K | Yes | Unspecified | |
| DeepSeek-V4-Pro-Max | 80.1 | 1.6T | 1M | 393.2K | No | MIT |
A concise description based on the published model registry.
Open MoE flagship with million-token context for coding and long agent runs
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-Pro.
DeepSeek-V4-Pro's default version was released on Apr 24, 2026.
No official standard PAYG price is currently available for DeepSeek-V4-Pro.
DeepSeek-V4-Pro 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 an unspecified license.
No published provider offering is currently linked to the default version.
Nearby ranked alternatives include Seed 2.1 Turbo, GLM 5.2, Hy3.