| Model profile | ||
|---|---|---|
| Model | Size | Approximate size |
| DeepSeek V4 ProModel source: provider1,048,576 tokens max native context | 1600Bactive 49B | 960 GB4-bit quantization |
| Compatible hardware Estimated speeds, not benchmark results: calculated from memory bandwidth and model size. Real results can differ significantly because there is no precise formula for deriving LLM generation speed from hardware specifications alone. | ||||
|---|---|---|---|---|
| Hardware | Memory | Bandwidth | vLLM | Estimated generation |
| DGX H200 | 1128 GBGPU HBM3e | 4800 GB/s | yes | 62 tok/s4-bit quantization |
| DGX B200 | 1440 GBGPU HBM3e | 8000 GB/s | yes | 97 tok/s4-bit quantization |
| GB200 NVL72 | 13400 GBGPU HBM3e | 8000 GB/s | yes | 97 tok/s4-bit quantization |
| GB300 NVL72 | 20000 GBGPU HBM3e | 8000 GB/s | yes | 97 tok/s4-bit quantization |
Ranks frontier models based on real use across front-end design, one-shot capabilities, cost, speed, and subscriptions, placing Fable 5 at the top.
Tests DeepSeek V4 Pro (0813) via WoAI Bench on frontend dev, agentic coding, 3D/Three.js, and one-shot gen, comparing cost, speed, and quality against Gemini 3.7 Flash, Grok 4.6, Kimi K3, GLM 5.2, Qwe
Evaluates DeepSeek V4 Pro and V4 Flash via BridgeMind horror game, Minecraft clone, Next.js dashboard, and BridgeMind 1 bug against GPT 5.6, Fable 5, and Grok 4.6 using DeepSeek Harness.
Tests DeepSeek V4 Pro across browser workflows, C++ game development, FPS generation, 3D CAD modeling, frontend design, and cinematic game creation.
Tests DeepSeek-V4-Pro-0813 via API and chat platforms, showing high performance in agent tests.