L LLM Cloud Hub
Self-host vs API

Gemma 2 27B Instruct

27B params · Gemma Terms of Use · google/gemma-2-27b-it

Cheapest API option for these weights
Google: Gemma 2 27B
$0.6500 in / $0.6500 out per 1M tokens
Monthly @ this workload
$487.50

Self-hosted scenarios

GPU Quant. Quantization — bf16 is full precision; fp8/awq-int4/gguf-q4 progressively shrink memory at slight quality cost. Glossary → tok/s Output tokens generated per second, single-stream. Batched serving achieves 5–10× higher aggregate. Glossary → $/hr GPU hours / mo Inference-only Best case — scale-to-zero serverless (RunPod Serverless, Modal). You pay only for the actual compute time you use. Glossary → Always-on (24/7) One dedicated GPU running all month (720h × $/hr). The floor for any sustained-load deployment. Glossary →
NVIDIA GeForce RTX 4090 24GB
RunPod (community cloud)
gguf-q4 35 0.3400 1190.5 $404.76 $244.80
NVIDIA RTX A6000 48GB
RunPod (community cloud)
bf16 50 0.4900 833.3 $408.33 $352.80
NVIDIA A100 80GB SXM
RunPod (community cloud)
bf16 70 1.7900 595.2 $1065.48 $1288.80

What these numbers mean. Inference-only assumes you can scale GPUs to zero between requests (RunPod Serverless, Modal, etc.) and only pay for the actual compute time — this is the floor. Always-on is the cost of one dedicated GPU running 24/7 (720 h/month) and is what most teams actually pay with a single VM. Throughput numbers are single-stream from public vLLM / TGI / llama.cpp benchmarks; with batched serving, real production deployments achieve 5–10× higher aggregate throughput. Use these as a starting point, not a quote.

Keyboard shortcuts

?
Show this overlay
/
Focus the first form field
g h
Go to / (home)
g b
Go to /best-llm-for
g c
Go to /cost
g s
Go to /self-hosted
g x
Go to /compliance
Esc
Close any overlay

Inspired by Linear and GitHub conventions. The two-key sequences (g then h) work within ~1 second.