DeepSeek-V4-Pro
DeepSeek·
LLMsopen-weightcloud + localFrontier
Overview
Preview release, available via API and open weights. 1.6T total / 49B active MoE with hybrid attention and manifold-constrained hyper-connections. Reports 27% of single-token inference FLOPs and 10% of KV cache vs DeepSeek-V3.2 in a 1M-token setting.
Capabilities and innovations
1.6T total / 49B active MoE1M token context windowThree reasoning effort modes (incl. Think Max)MIT licenseHybrid attention: CSA + HCAManifold-Constrained Hyper-Connections27% inference FLOPs and 10% KV cache vs V3.2 at 1M tokens
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 90.1% | vendor Official model card, instruct table, Think Max column: GPQA Diamond 90.1 (High mode 89.1, Non-Think 72.9). |
| Humanity’s Last Exam (no tools) | 37.7% | vendor Official model card, Think Max column: HLE 37.7. The card does not state whether tools were enabled. |
| SWE-bench Verified | 80.6% | vendor Official model card, Think Max column: SWE-bench Verified 80.6. |
| SWE-bench Pro | 55.4% | vendor Official model card, Think Max column: SWE-bench Pro 55.4. |
| MMLU-Pro | 87.5% | vendor Official model card, Think Max column: MMLU-Pro 87.5. |
| Terminal-Bench 2.x | 50.19% | independent Terminal-Bench 2.1 50.19%, measured by Vals AI with the Terminus 2 harness (accessed 2026-09-23). Vals AI is the reference harness for this field so values compare across vendors; vendor and other third-party figures stay in the notes below. Listed by Vals AI as "DeepSeek V4" at V4-Pro pricing; the later 0813 build scores 54.68 on the same harness. Replaces 67.9 (vendor). |
- MMMU-Pro: not reported by the vendor
API pricing
$0.44 per million input tokens, $0.87 per million output tokens (USD, cheapest OpenRouter endpoint, checked 2026-07-31)
Architecture and hardware
- Parameters
- 1600B total, 49B active per token (MoE)
- Estimated VRAM at Q4
- ~920 GB, Frontier class
Reliability
- Hallucination rate (Vectara HHEM)
- 8.6% (lower is better)independent
Links
More from DeepSeek
Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.