DeepSeek · Chat and reasoning · 1.6T parameters · 49B active · mit
DeepSeek V4 Pro 0813
Needs 1,072 GB at INT4 / MXFP4. The smallest real configuration today is 8× MI300X at TensorWave, $13.68 per hour, about $9,986 a month at 730 hours.
Parameters1.6T49B active per token
Context1M tokensmaximum published
Checkpoint893 GBINT4-packed, as published
Licensemitcheck before commercial use
ReleasedAug 13, 2026documented date
Smallest configuration · INT4 / MXFP4
8× MI300X1536 GB VRAM · 70% used
- Per hour
- $13.68
- Per month
- $9,986
- Cheapest at
TensorWave
Memory · 1 precision · 6 context lengthsHow much VRAM it needs
Weights plus runtime overhead, then the KV cache that grows with the context you serve.
| Precision | Weights | Weights + headroom | 4k context | 32k context | 128k context | 256k context | 1M context |
|---|---|---|---|---|---|---|---|
| INT4-packedpublished checkpoint | 893 GB | 1,072 GB | 1,072 GB | 1,076 GB | 1,088 GB | 1,105 GB | 1,203 GB |
| KV cacheone sequence, FP16 cache, included above | — | — | 1 GB | 5 GB | 17 GB | 33 GB | 131 GB |
Cross-checked on Hugging Face Oct 28, 2026: 1650.5B parameters in the safetensors headers, 892.7 GB of weight files, attention layout from config.json (61 layers, 1 KV heads × 512). Method: weights as published at the native precision, verified on Hugging Face 2026-10-28; other precisions estimated at 1 (8-bit) and 0.6 (4-bit) bytes per parameter; × 1.2 runtime headroom; KV cache = 2048 bytes per token per layer × 61 full-attention layers (FP16 cache, one sequence).config.json ↗
Smallest configuration per GPU · 109 GPUs · cheapest eligible on-demand offer per GPU count, every provider on fileWhat it costs to run, GPU by GPU
The fewest GPUs of one type whose total memory fits the model at short context, priced from the cheapest eligible on-demand offer for exactly that count (45-day window, newest reading per provider), with the next providers beside it. Monthly = 730 hours.
| GPU | Smallest config | VRAM used | $ / hour | $ / month | Cheapest at |
|---|---|---|---|---|---|
| MI300XBest fit | 8× 192 GB1536 GB total | 70% | $13.68 | $9,986 | TensorWaveTrust 64 · read Nov 2 · receipt · also Cyfuture AI $22.38, AtmosCompute $26.77 |
| MI325X | 8× 256 GB2048 GB total | 52% | $18.00 | $13,140 | TensorWaveTrust 64 · read Nov 2 · receipt · also Cyfuture AI $24.34 |
| H200 | 8× 141 GB1128 GB total | 95% tight | $21.04 | $15,359 |  Airon AInot scored · read Nov 2 · receipt · also Geodd $21.76, Lium $23.92 |
| B300 | 4× 270 GB1080 GB total | 99% tight | $23.29 | $17,004 |  UpCloudTrust 75 · read Nov 2 · receipt · also FPT AI Factory $27.96, Verda $45.40 |
| MI355X | 8× 288 GB2304 GB total | 46% | $23.60 | $17,228 | TensorWaveTrust 64 · read Nov 2 · receipt |
| H200 NVL | 8× 141 GB1128 GB total | 95% tight | $26.09 | $19,042 |  IONOSnot scored · read Nov 2 · receipt · also Geodd $27.12, Vast.ai $39.47 |
| B200 | 8× 180 GB1440 GB total | 74% | $32.91 | $24,026 | UpCloudTrust 75 · read Nov 2 · receipt · also Airon AI $33.04, Deep Infra $39.12 |
| B300 CC | 4× 270 GB1080 GB total | 99% tight | $46.31 | $33,806 |  VerdaTrust 81 · read Nov 2 · receipt |
| GB300 | 4× 279 GB1116 GB total | 96% tight | $52.21 | $38,113 | VerdaTrust 81 · read Nov 2 · receipt |
| B200 CC | 8× 180 GB1440 GB total | 74% | $72.37 | $52,830 | VerdaTrust 81 · read Nov 2 · receipt |
Throughput estimates · Oct 4, 2026Tokens per second, and what a million costs
Chat and reasoning. Speed estimated from memory bandwidth and model size, never measured. The cost per 1M output tokens is an estimate at the prices read on Oct 4, 2026; it is not recomputed here.
| GPU | Cards | Precision | tok/s per user | per 1M output tokens · estimate |
|---|---|---|---|---|
| MI300X | 8 | FP4 | 34.5 | $4.36 |
| B200 | 8 | FP4 | 59.8 | $5.29 |
| H200 | 8 | FP4 | 44.8 | $5.42 |
cards, precision and speed estimated from memory bandwidth and model size; cost recomputed by the site from verified prices · cheapest per output at the prices of Oct 4, 2026: MI300XWhy estimates, and their limits →
DeepSeek V4 Pro 0813 · related