Blog

Run AI on GPUs

Hands-on guides, product updates and engineering notes.
Get new posts by e-mail: Subscribe to the Lium newsletter
Latest
· 4 min read

Best GPU cloud for fine-tuning a 70B model in 2026: memory needs, node choice and cost per hour

QLoRA on a 70B model fits one 96 GB GPU ($1.49 per hour on Lium Secure), LoRA on a 16-bit base needs about 140 GB (one B200, $7.00 per hour), and a full fine-tune needs about 1.26 TB before activations (8x B300, $32 per hour on Spot). Prices checked 15 November 2026.
fine-tuning
training
qlora
lora
b200
b300
buyer-guide

More articles

· 4 min read

NVIDIA B300 rental price per hour in 2026: Spot vs Secure on Lium, and when B200 or H200 costs less

On Lium a whole 8x B300 node costs $4.00 per GPU-hour on Spot ($32 per hour) and from $9.00 on Secure; one B300 on Secure starts at $7.14. Per hour, 8x B200 costs less, and one H200 at $3.80 costs about half a single Secure B300; B300 is cheaper per job only if it runs your job faster by the break-even factor. Prices checked 15 November 2026.
gpu-pricing
b300
b200
h200
blackwell
buyer-guide
· 3 min read

RTX 5090 cloud rental in 2026: price per hour, what fits in 32 GB, and when it beats a 4090 or an H100

Rent one RTX 5090 on Lium from $0.55 per hour on Spot or $0.65 on Secure, or a whole 8-GPU node from $0.49 per GPU-hour. 32 GB holds an 8B model in 16-bit or a 32B model in 4-bit. Prices checked 15 November 2026.
rtx-5090
inference
pricing
buyer-guide
· 6 min read

Cost per million tokens: self-hosting Qwen3.8 and Nemotron 3.5 on a rented H100 vs paying per token

We measured Qwen3.8-27B and Nemotron 3.5 Lightning on one $1.40/hour H100 on Lium. With the GPU kept busy, it did $12.37 an hour of Qwen3.8-27B API work for $1.40, and served Nemotron at $0.14 per million output tokens.
llm-inference
cost
vllm
qwen
nemotron
h100
pricing
· 5 min read

B200 vs H200 vs H100 for LLM training in 2026: which GPU to rent

On the Lium Secure tier an 8x H100 node costs $2.90 per GPU-hour and 8x H200 $4.06; on Spot, 8x H100 costs $1.49 and 8x B200 $5.50. On Spot, a B200 only saves money if your job runs at least 3.7x faster than on H100. Prices checked 9 November 2026.
gpu-pricing
training
b200
h200
h100
comparison
· 2 min read

How Seinari benchmarked on up to 16 RTX 5090s at once on Lium

Seinari ran 120 RTX 5090 pods on Lium, up to 16 at once, and paid $0.72 to $0.79 per GPU-hour on average. Runpod lists the same card at $1.19 on Secure Cloud.
case-study
rtx-5090
inference
benchmark
· 10 min read

Lium vs AWS and Google Cloud: H100 training at less than half the on-demand price

One 8x H100 SXM node on Lium costs $556.80 for 24 hours vs $1,320.96 on AWS and $2,123.76 on Google Cloud, with no quota request and agent budgets built in.
comparison
aws
google-cloud
h100
training
pricing
· 3 min read

GPU cloud MCP servers compared

We checked official GPU cloud MCP servers from six providers. Only Lium's hosted server returned live GPU prices without an account in our checks. Lium also documents daily, monthly and lifetime USD budgets per agent key.
comparison
mcp
ai-agents
claude-code
codex
· 8 min read

Lium vs Lambda: same GPUs, lower prices, built for AI agents

We compared on-demand H100, H200, B200 and A6000 prices, billing, storage and AI agent support across Lium and Lambda on 29 October 2026.
comparison
lambda
h100
b200
pricing
ai-agents
· 6 min read

Best GPU cloud providers in 2026

We compared on-demand H100 SXM, B200 and B300 prices across Lium, RunPod, Vast.ai, Lambda, CoreWeave, Nebius and Together AI. Lium had the lowest H100 SXM and B200 price for a whole 8-GPU node.
comparison
gpu-pricing
h100
b200
b300
· 1 min read

Case study: how Ahurasense runs B300 workloads on Lium

Ahurasense puts Lium B300 nodes behind its own AI GPU pods. The nodes are about 99% stable and it gets as many instances as it needs.
case-study
b300
reliability
· 7 min read

Lium vs Vast.ai: where should you rent a GPU?

Compare Lium and Vast.ai on GPU prices, spending controls for coding agents, host checks and failed-start protections, then bring your existing job to Lium.
comparison
vast-ai
h100
h200
pricing
· 6 min read

Cheapest H100 and H200 cloud in 2026: 8-hour prices

We priced one 8-hour job on an 8x H200 node, an 8x H100 node and one H100 across Lium, RunPod, Vast.ai, Lambda and AWS. Every price is dated and sourced.
gpu-pricing
h100
h200
comparison
· 4 min read

Give Claude Code a GPU in 60 seconds

Let Claude Code rent a Lium GPU, run code on it, bring results back and remove it, with a scoped API key, a server-enforced daily budget and automatic removal. Tested on lium.io; the whole run cost under one cent.
claude-code
ai-agents
cli
rtx-5070-ti
· 3 min read

Give OpenAI Codex a GPU with the Lium CLI

Let the Codex CLI rent a Lium GPU, run your code on it and clean up after itself, using an AGENTS.md rules block, a scoped API key with a daily budget, and automatic pod removal.
codex
ai-agents
cli
· 8 min read

Lium vs RunPod: why developers and AI agents rent their H100s on Lium

Lium beat RunPod on price for every GPU in our 17 October 2026 table, charges nothing for a pod that fails to start, and lets Claude Code or Codex rent a GPU on its own. Here is where Lium wins, with the numbers.
comparison
runpod
h100
a100
pricing
ai-agents
· 5 min read

Run ComfyUI on a Lium GPU

Install ComfyUI on a Lium RTX 4090, load the openly licensed SDXL base checkpoint, open the UI through an SSH tunnel and generate an image through the ComfyUI HTTP API. 19 minutes of pod time, about 15 cents.
comfyui
stable-diffusion
sdxl
image-generation
rtx-4090
· 6 min read

Fine-tune an open LLM with LoRA on Lium

Fine-tune Qwen2.5-0.5B-Instruct with LoRA on a Lium GPU using Hugging Face transformers and PEFT, then copy the adapter back. Five minutes of pod time, about three cents.
lora
fine-tuning
llm
qwen
peft
rtx-a6000
· 5 min read

Serve an open LLM with vLLM on a single Lium GPU

Serve Qwen2.5-7B-Instruct with vLLM on one RTX 4090 on Lium and call your own OpenAI-compatible endpoint from your laptop with curl and the openai Python client. 18 minutes of pod time, about 14 cents.
vllm
llm-inference
qwen
openai-api
rtx-4090
· 5 min read

Train a PyTorch model on a Lium GPU

Rent a GPU on Lium, train ResNet-18 on CIFAR-10 in the background, follow the log, copy the checkpoint back and remove the pod. About ten cents on an RTX A6000.
pytorch
training
resnet
cifar-10
rtx-a6000
· 7 min read

Best GPU cloud for open-source LLM inference in 2026

How to choose a GPU cloud for serving Llama, Qwen and other open-weight LLMs, which GPU each model size needs, and what it costs on Lium at live prices.
llm-inference
vllm
llama
h200
gpu-pricing
· 5 min read

Run DeepSeek V4.1 Flash on your own Lium GPUs

Run DeepSeek-V4.1-Flash on a Lium GPU node with vLLM, then call your own OpenAI-compatible endpoint from your laptop.
deepseek
vllm
llm-inference
h100