NVIDIA48 GB VRAM960 GB/s$0.8/hr

GPT-OSS 20B (MoE) on RTX 6000 Ada 48GB: Does It Fit, How Fast, What It Costs

GPT-OSS 20B (MoE) at FP8 (E4M3) needs 20.9 GB for its weights and 22.6 GB in total at 8K context with 8 concurrent users. That means it fits on a single RTX 6000 Ada 48GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
22.6 GB needed / 48 GB available
Decode speed
260.0
tok/s per stream
Aggregate
3120
tok/s @ 8 concurrent
Cost / 1M tokens
$0.0712
1× RTX 6000 Ada 48GB
What changes as you add RTX 6000 Ada 48GBs
Tensor-parallel degree vs VRAM, throughput, and cost. Cost scales linearly with GPU count; decode throughput scales near-linearly too, so cost per token stays roughly flat.
GPUsTotal VRAMVRAM neededDecode tok/sAggregate tok/sTTFTCost / 1M tokFits?
1× RTX48 GB22.6 GB260.031201.32 s$0.0712✓
2× RTX96 GB22.6 GB442.010608676.2 ms$0.0419✓
4× RTX192 GB22.6 GB884.042432353.1 ms$0.0209✓
8× RTX384 GB22.6 GB1768169728191.5 ms$0.0105✓
Concurrency vs context length
On 1× RTX 6000 Ada 48GB at FP8 (E4M3). KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K260206.2 MB✓ fits
8K132412.3 MB✓ fits
33K331.61 GB✓ fits
66K163.22 GB✓ fits
131K86.44 GB✓ fits
Other GPUs that run GPT-OSS 20B (MoE)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB1×5824$0.0238
RTX 3090 24GB1×2472$0.0281
RTX 4090 24GB1×3276$0.0339
AMD MI300X 192GB1×17225$0.0403
V100 SXM2 32GB1×2925$0.0475
A100 40GB SXM41×5038$0.0496
Frequently asked questions

Can GPT-OSS 20B (MoE) run on RTX 6000 Ada 48GB?

Yes. GPT-OSS 20B (MoE) at FP8 (E4M3) needs 22.6 GB including the KV cache, and RTX 6000 Ada 48GB has 48 GB — so it fits on a single GPU.

How fast is GPT-OSS 20B (MoE) on RTX 6000 Ada 48GB?

260.0 tokens/sec per stream and 3120 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 1.32 s, and inter-token latency is 3.8 ms.

What does GPT-OSS 20B (MoE) on RTX 6000 Ada 48GB cost per million tokens?

$0.0712 per 1M output tokens at 1× RTX 6000 Ada 48GB ($0.8000/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does GPT-OSS 20B (MoE) need on RTX 6000 Ada 48GB?

20.9 GB for the weights at FP8 (E4M3), plus 1.6 GB of KV cache for 8 concurrent requests at 8K context — 22.6 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is RTX 6000 Ada 48GB the right choice for GPT-OSS 20B (MoE)?

It works. Compare $0.0712 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 4× RTX 3080 10GB at $0.006615, and the fastest is 1× Cerebras CS-3 (WSE-3) at 65000 tok/s.

Adjust the workload yourself

Same formulas, your context length, concurrency, and quantisation.

Open in calculator