MODEL ARCHITECTURE SPEC • DEEPSEEK236B MoE
DeepSeek Coder V2 236B MoE VRAM & inference speed
236B Total Params16B Active (MoE)60 Layers • 8 KV Heads32K Default Ctx (Max 125K)Ollama: deepseek-coder-v2:236b
Accurate memory requirements, KV cache footprint, and tokens/sec generation benchmarks for running DeepSeek Coder V2 236B MoE across local GPUs, Apple Silicon Macs, and cloud instances.
✓ FITS IN VRAM151.34 GB (64%)~54 t/s
DeepSeek Coder V2 236B MoE on Apple Mac Studio M5 Ultra (256GB Unified Memory)
✓ Fits Comfortably in VRAM
✓ FITS IN VRAM
Required VRAM
151.34 GB
64% of hardware capacity (236 GB)Est. Decode Speed
~54 tok/s
16B active params @ 890 GB/sAvailable Headroom
+84.66 GB
Safe for OS display & full contextMemory Allocation & Headroom
151.34 GB allocated of 236 GB capacity
64% utilized
Weights: 141.6 GB
KV: (32K): 4.03 GB
Overhead: 5.71 GB
Free Headroom: +84.66 GB
⚡ Architecture & Throughput Transparency: Single-user decode speed (~54 tok/s) is memory-bandwidth bound at batch size = 1 on Apple Mac Studio M5 Ultra (256GB Unified Memory) (890 GB/s). Prompt prefill / TTFT (~783 tok/s) is compute-bound and utilizes GPU Tensor cores. Actual latency varies with FlashAttention-2 and runtime engine (Ollama/vLLM/llama.cpp).
OLLAMA_FLASH_ATTENTION=1 OLLAMA_KV_CACHE_TYPE=q8_0 ollama run deepseek-coder-v2:236bCloud Instance Options for DeepSeek Coder V2 236B MoE
On-DemandDeploy this model on an on-demand cloud GPU without local hardware constraints or power usage:
RunPod Cloud GPUsBest 1-Click Ollama & vLLM
$0.34/hr (RTX 4070 Ti) – $2.39/hr (H100)Vast.ai GPU MarketplaceCheapest Dual RTX 3090 / 4090
$0.18/hr (RTX 3060) – $0.55/hr (RTX 4090)Lambda GPU CloudEnterprise 8x H100 Clusters
$1.99/hr (GH200 / H100)Quantization Precision Matrix
32K ctx • KV: Q8_0| Quantization | Weights | KV Cache | Total VRAM | Est. Speed | Fit Status |
|---|---|---|---|---|---|
| FP1616 bpw | 472 GB | 4.03 GB | 493.3 GB (209%) | 0 t/s | OOM |
| Q8_08.5 bpw | 250.16 GB | 4.03 GB | 263.7 GB (112%) | ~7 t/s | PARTIAL |
| Q6_K6.56 bpw | 193.52 GB | 4.03 GB | 205.07 GB (87%) | ~40 t/s | FITS |
| Q5_K_M5.68 bpw | 167.56 GB | 4.03 GB | 178.2 GB (76%) | ~46 t/s | FITS |
| Q4_K_M4.8 bpw | 141.6 GB | 4.03 GB | 151.34 GB (64%) | ~54 t/s | FITS |
| FP44.25 bpw | 125.08 GB | 4.03 GB | 134.24 GB (57%) | ~60 t/s | FITS |
| IQ3_XS3.3 bpw | 96.76 GB | 4.03 GB | 104.93 GB (44%) | ~75 t/s | FITS |
Export Full Architecture & Infrastructure Spec
Get an executive-ready PDF report with exact memory math, hardware benchmarks, and cost comparisons.
Hardware Compatibility Matrix for DeepSeek Coder V2 236B MoE
Standard Q4_K_M precision with 32K context window and Q8_0 KV cache
| Hardware System | VRAM Envelope | Required VRAM | Headroom | Est. Decode Speed | Compatibility | Test Fit |
|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5090 (32GB GDDR7) Consumer GPU • 1792 GB/s | 32 GB | 151.34 GB | -119.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 4090 (24GB GDDR6X) Consumer GPU • 1008 GB/s | 24 GB | 151.34 GB | -127.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| 2x NVIDIA RTX 3090 Dual Rig (48GB Total VRAM) Multi-GPU Rig • 936 GB/s | 48 GB | 151.34 GB | -103.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 5080 (16GB GDDR7) Consumer GPU • 960 GB/s | 16 GB | 151.34 GB | -135.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA RTX 4070 Ti Super (16GB) Consumer GPU • 672 GB/s | 16 GB | 151.34 GB | -135.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 3060 (12GB) Consumer GPU • 360 GB/s | 12 GB | 151.34 GB | -139.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| Apple Mac Mini M4 Pro (64GB Unified Memory) Apple Silicon • 273 GB/s | 54 GB | 151.34 GB | -97.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| Apple Mac Studio M4 Max (128GB Unified Memory) Apple Silicon • 546 GB/s | 112 GB | 151.34 GB | -39.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| Apple Mac Studio M5 Ultra (256GB Unified Memory) Apple Silicon • 890 GB/s | 236 GB | 151.34 GB | +84.66 GB | ~54 tok/s | ✓ FITS | Test Fit → |
| AMD Strix Halo Mini-PC (Ryzen AI Max+ 395 128GB) Unified APU / Mini-PC • 256 GB/s | 110 GB | 151.34 GB | -41.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA DGX Spark (128GB Grace-Blackwell Unified) AI Appliance • 273 GB/s | 118 GB | 151.34 GB | -33.34 GB | ~4 tok/s | ⚡ PARTIAL | Test Fit → |
| NVIDIA H100 SXM (80GB HBM3 Cloud GPU) Datacenter Cloud GPU • 3350 GB/s | 80 GB | 151.34 GB | -71.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 5070 Ti (16GB GDDR7) Consumer GPU • 896 GB/s | 16 GB | 151.34 GB | -135.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 5070 (12GB GDDR7) Consumer GPU • 672 GB/s | 12 GB | 151.34 GB | -139.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| NVIDIA GeForce RTX 4080 Super (16GB GDDR6X) Consumer GPU • 736 GB/s | 16 GB | 151.34 GB | -135.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| AMD Radeon RX 7900 XTX (24GB GDDR6) Consumer GPU • 960 GB/s | 24 GB | 151.34 GB | -127.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| AMD Radeon RX 7900 XT (20GB GDDR6) Consumer GPU • 800 GB/s | 20 GB | 151.34 GB | -131.34 GB | 0 (OOM) | ✕ OOM | Test Fit → |
| Apple Mac Studio M2 Ultra (192GB Unified Memory) Apple Silicon Workstation • 800 GB/s | 144 GB | 151.34 GB | -7.34 GB | ~11 tok/s | ⚡ PARTIAL | Test Fit → |