MiniMax M2.5 (229B-A10B) VRAM Requirements, KV Cache & Best GPUs
MiniMax M2.5 (229B-A10B) (228.69B total, 10B active, Sparse MoE + GQA) uses 128.84 GiB for 4-bit weights (Q4_K_M), requiring 131.34 GiB at 8K context and 137.15 GiB at 32K. · Hugging Face: MiniMaxAI/MiniMax-M2.5 ↗ · GGUF: unsloth/MiniMax-M2.5-GGUF ↗
Fits with Safe Headroom
Measured GGUF shard size
HF config.json ↗
Required GPU Memory
131.34 GiB
141.03 GB (decimal)
Usable Hardware Memory
182.4 GiB
192 GB (physical)
Safe Headroom
+51.06 GiB
68% utilization
Est. Decode Throughput
~218 tok/s
10B active params
Est. Prefill Throughput
~24000 tok/s
4K Prompt TTFT: ~171 ms
Est. Cost / 1M Tokens
$3.30
Sparse MoE + GQA · Custom
Recommended Hardware for Current Configuration
Best Single GPU
Instinct MI300X 192GB (192GB)
~$15,000
Best Value / Used GPU
2x A100 80GB SXM
160GB · ~$24,000
Apple Unified Memory
Mac Studio M4 Ultra 256GB
~$5,999
Cloud From
DigitalOcean · $2.59 / hr
1x Instinct MI300X 192GB
Production Launch Commands
llama.cpp (GGUF)
llama-server -hf unsloth/MiniMax-M2.5-GGUF:Q4_K_M \ --ctx-size 8192 -ngl 999 --flash-attn
vLLM
vllm serve MiniMaxAI/MiniMax-M2.5 \ --max-model-len 8192 --gpu-memory-utilization 0.92
Live Token Speed Simulator (TTFT ~171ms · ~218 tok/s)
[MiniMax M2.5 (229B-A10B) · 131.34 GiB] allocated on target GPU. Prefill completed in TTFT window; streaming autoregressive tokens at memory-bandwidth-bound rate...
Cloud GPU Comparison & Rent-vs-Buy TCO
Verified public rates across RunPod, Vast.ai, Vultr, and DigitalOcean.
Purchase break-even
~21.1 months
Est. local power: $15.96 / mo
| Provider | Recommended Instance | Total VRAM | Hourly Rate | Monthly at Selected Hours | Billing Mode | Deploy |
|---|---|---|---|---|---|---|
| RunPod | 1x B200 180GB (Community Cloud) | 180 GB | $5.98 / hr | $727.17 / mo | Per-second | Deploy GPU ↗ |
| Vast.ai | 1x B200 180GB (Verified Host) | 180 GB | $7.50 / hr | $912.24 / mo | Spot / On-demand | Deploy GPU ↗ |
| Vultr ($300 Free Credit*) | 4x L40S 48GB (Dedicated Instance · $300 Credit*) | 192 GB | $6.68 / hr | $812.77 / mo | Hourly | Deploy GPU ↗ |
| Vultr Cloud GPU | 4x L40S 48GB (Dedicated Instance) | 192 GB | $6.68 / hr | $812.77 / mo | Hourly | Deploy GPU ↗ |
| DigitalOcean | 1x Instinct MI300X 192GB (Dedicated Instance) | 192 GB | $2.59 / hr | $314.94 / mo | Per-second | Deploy GPU ↗ |
Rates reflect public listings on the verification date and fluctuate with regional inventory.
Fine-Tuning VRAM Requirements (QLoRA / LoRA / Full) — MiniMax M2.5 (229B-A10B)
QLoRA (4b)
Fits with Safe Headroom
131.28 GiB
Model Weights: 128.84 GiB · Gradients + Optimizer: 1.4 GiB · Act: 1.04 GiB
Recommended GPU: Instinct MI300X 192GB (192GB)
LoRA (16b)
Exceeds Available GPU Memory
429.66 GiB
Model Weights: 426.07 GiB · Gradients + Optimizer: 2.2 GiB · Act: 1.39 GiB
Recommended GPU: 4x H200 141GB SXM (564GB)
Full BF16
Exceeds Available GPU Memory
1749.33 GiB
Model Weights: 426.07 GiB · Gradients + Optimizer: 1320.82 GiB · Act: 2.44 GiB
Recommended GPU: Multi-GPU Required
Engineering Sizing Notes & Optimization Playbook
Quick Hardware Verdict
Instinct MI300X 192GB (192GB physical, 182.4 GiB usable) runs 4-bit (Q4_K_M) MiniMax M2.5 (229B-A10B) at 8K context using 131.34 GiB with 51.06 GiB of safe headroom (~218 tok/s single-stream decode).
KV Cache & Runtime Tuning
MiniMax M2.5 (229B-A10B) uses Sparse MoE + GQA (228.69B total, 10B active per token). Moving from 8K to 32K context expands FP16 KV cache from 1.94 GiB to 7.75 GiB; enabling 8-bit KV cache (llama.cpp -ctk q8_0 -ctv q8_0 or vLLM --kv-cache-dtype fp8) saves 3.87 GiB at 32K (reducing total VRAM to 133.28 GiB).
VRAM Scaling Across Context Windows (4K to 128K, 4-bit Weights)
MiniMax M2.5 (229B-A10B) × 49 GPUs (4-bit Q4_K_M)
| Target Hardware | Total VRAM | Memory Bandwidth | 4-bit @ 8K | 4-bit @ 32K | Decode Speed | Compatibility | Full Compatibility Report |
|---|---|---|---|---|---|---|---|
| NVIDIA RTX 3060 12GB | 12 GB (11.4 GiB) | 360 GB/s | 131.34 GiB | 137.15 GiB | ~8.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 4060 Ti 16GB | 16 GB (15.2 GiB) | 288 GB/s | 131.34 GiB | 137.15 GiB | ~8.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 4070 Ti Super | 16 GB (15.2 GiB) | 672 GB/s | 131.34 GiB | 137.15 GiB | ~8.4 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 4080 Super | 16 GB (15.2 GiB) | 736 GB/s | 131.34 GiB | 137.15 GiB | ~8.5 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 5070 12GB | 12 GB (11.4 GiB) | 672 GB/s | 131.34 GiB | 137.15 GiB | ~8.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 5070 Ti | 16 GB (15.2 GiB) | 896 GB/s | 131.34 GiB | 137.15 GiB | ~8.5 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 5080 | 16 GB (15.2 GiB) | 960 GB/s | 131.34 GiB | 137.15 GiB | ~8.5 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 3090 | 24 GB (22.8 GiB) | 936 GB/s | 131.34 GiB | 137.15 GiB | ~9 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 4090 | 24 GB (22.8 GiB) | 1008 GB/s | 131.34 GiB | 137.15 GiB | ~9.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 4090 D | 24 GB (22.8 GiB) | 1008 GB/s | 131.34 GiB | 137.15 GiB | ~9.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 5090 | 32 GB (30.4 GiB) | 1792 GB/s | 131.34 GiB | 137.15 GiB | ~9.8 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 5090 D | 32 GB (30.4 GiB) | 1792 GB/s | 131.34 GiB | 137.15 GiB | ~9.8 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX A6000 | 48 GB (45.6 GiB) | 768 GB/s | 131.34 GiB | 137.15 GiB | ~11 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX 6000 Ada | 48 GB (45.6 GiB) | 960 GB/s | 131.34 GiB | 137.15 GiB | ~11.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA RTX PRO 6000 Blackwell | 96 GB (91.2 GiB) | 1792 GB/s | 131.34 GiB | 137.15 GiB | ~23.7 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA DGX Spark (GB10 128GB) | 128 GB (99.8 GiB) | 273 GB/s | 131.34 GiB | 137.15 GiB | ~21.4 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA L4 24GB | 24 GB (22.8 GiB) | 300 GB/s | 131.34 GiB | 137.15 GiB | ~8.6 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA A40 48GB | 48 GB (45.6 GiB) | 696 GB/s | 131.34 GiB | 137.15 GiB | ~11 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA L20 48GB | 48 GB (45.6 GiB) | 864 GB/s | 131.34 GiB | 137.15 GiB | ~11.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA L40S 48GB | 48 GB (45.6 GiB) | 864 GB/s | 131.34 GiB | 137.15 GiB | ~11.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA A800 80GB SXM | 80 GB (76 GiB) | 2039 GB/s | 131.34 GiB | 137.15 GiB | ~17.6 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA A100 80GB SXM | 80 GB (76 GiB) | 2039 GB/s | 131.34 GiB | 137.15 GiB | ~17.6 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA H800 80GB SXM | 80 GB (76 GiB) | 3350 GB/s | 131.34 GiB | 137.15 GiB | ~18.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA H100 80GB SXM | 80 GB (76 GiB) | 3350 GB/s | 131.34 GiB | 137.15 GiB | ~18.1 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA H20 96GB HBM3 | 96 GB (91.2 GiB) | 4000 GB/s | 131.34 GiB | 137.15 GiB | ~25.4 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| NVIDIA H20 141GB HBM3e | 141 GB (133.9 GiB) | 4800 GB/s | 131.34 GiB | 137.15 GiB | ~206.8 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| NVIDIA H200 141GB SXM | 141 GB (133.9 GiB) | 4800 GB/s | 131.34 GiB | 137.15 GiB | ~206.8 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| NVIDIA B200 180GB | 180 GB (171 GiB) | 8000 GB/s | 131.34 GiB | 137.15 GiB | ~264.4 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| NVIDIA B300 288GB Blackwell Ultra | 288 GB (273.6 GiB) | 8000 GB/s | 131.34 GiB | 137.15 GiB | ~264.4 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| Huawei Ascend 910B 64GB (昇腾) | 64 GB (60.8 GiB) | 1200 GB/s | 131.34 GiB | 137.15 GiB | ~13.5 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Huawei Ascend 910C 128GB (昇腾) | 128 GB (121.6 GiB) | 2400 GB/s | 131.34 GiB | 137.15 GiB | ~94 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Huawei Atlas 300I Duo 96GB (310P) | 96 GB (91.2 GiB) | 408 GB/s | 131.34 GiB | 137.15 GiB | ~16.7 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Cambricon MLU590 80GB (寒武纪思元) | 80 GB (76 GiB) | 2048 GB/s | 131.34 GiB | 137.15 GiB | ~17.7 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Hygon DCU K100-AI 64GB (海光深算) | 64 GB (60.8 GiB) | 1638 GB/s | 131.34 GiB | 137.15 GiB | ~13.8 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| MetaX MetaX C500 64GB (沐曦曦云) | 64 GB (60.8 GiB) | 1800 GB/s | 131.34 GiB | 137.15 GiB | ~13.9 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Kunlunxin Kunlunxin P800 64GB (昆仑芯) | 64 GB (60.8 GiB) | 1600 GB/s | 131.34 GiB | 137.15 GiB | ~13.8 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Moore Threads MTT S4000 48GB (摩尔线程) | 48 GB (45.6 GiB) | 768 GB/s | 131.34 GiB | 137.15 GiB | ~11 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| AMD Radeon RX 7900 XTX 24GB | 24 GB (22.8 GiB) | 960 GB/s | 131.34 GiB | 137.15 GiB | ~9 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| AMD Radeon AI PRO R9700 32GB | 32 GB (30.4 GiB) | 640 GB/s | 131.34 GiB | 137.15 GiB | ~9.5 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| AMD Ryzen AI Max+ 395 128GB | 128 GB (99.8 GiB) | 256 GB/s | 131.34 GiB | 137.15 GiB | ~20.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| AMD Instinct MI300X 192GB | 192 GB (182.4 GiB) | 5300 GB/s | 131.34 GiB | 137.15 GiB | ~218 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| AMD Instinct MI325X 256GB | 256 GB (243.2 GiB) | 6000 GB/s | 131.34 GiB | 137.15 GiB | ~232.1 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| AMD Instinct MI355X 288GB | 288 GB (273.6 GiB) | 8000 GB/s | 131.34 GiB | 137.15 GiB | ~264.4 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| Intel Arc Pro B60 24GB | 24 GB (22.8 GiB) | 456 GB/s | 131.34 GiB | 137.15 GiB | ~8.8 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Intel Gaudi 3 128GB HBM2e | 128 GB (121.6 GiB) | 3700 GB/s | 131.34 GiB | 137.15 GiB | ~113.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Apple Mac mini M4 Pro 64GB | 64 GB (49.9 GiB) | 273 GB/s | 131.34 GiB | 137.15 GiB | ~11.2 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Apple MacBook Pro M4 Max 128GB | 128 GB (99.8 GiB) | 546 GB/s | 131.34 GiB | 137.15 GiB | ~38.3 tok/s | Exceeds Available GPU Memory | Full Compatibility Report → |
| Apple Mac Studio M4 Ultra 256GB | 256 GB (199.7 GiB) | 819 GB/s | 131.34 GiB | 137.15 GiB | ~61.4 tok/s | Fits with Safe Headroom | Full Compatibility Report → |
| Apple Mac Studio M3 Ultra 512GB | 512 GB (399.4 GiB) | 819 GB/s | 131.34 GiB | 137.15 GiB | ~61.4 tok/s | Fits with Safe Headroom | Full Compatibility Report → |