https://peer.madiator.cloud/videos/watch/67731369-4b43-476c-a14d-f15edf8a93b7 Spot instances on Runpod can provide considerable savings over on-demand, with the tradeoff that your instance can be interrupted depending on customer demand and other factors. This means that certain use cases will be preferred for spot; primarily batch work such as bulk image, video, or text processing or checkpointed training that can be resumed with a minimal amount of effort. REST API example used: #!/bin/bash Runpod Spot Instance Creation CommandReplace variables with your actual valuesSet your variablesAPI_TOKEN="your_runpod_api_token_here" TEMPLATE_ID="your_template_id" # Optional - leave null if not using a template GPU_TYPE="NVIDIA GeForce RTX 4090" # Examples: "NVIDIA A100 80GB PCIe", "NVIDIA GeForce RTX 4090", "NVIDIA RTX A6000" GPU_COUNT=1 VOLUME_SIZE=20 # Volume size in GB CONTAINER_SIZE=50 # Container disk size in GB POD_NAME="my-spot-instance" Create the spot instancecurl --request POST --url https://rest.runpod.io/v1/pods --header "Authorization: Bearer ${API_TOKEN}" --header "Content-Type: application/json" --data '{ "name": "'${POD_NAME}'", "imageName": "runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04", "gpuCount": '${GPU_COUNT}', "gpuTypeIds": ["'${GPU_TYPE}'"], "volumeInGb": '${VOLUME_SIZE}', "volumeMountPath": "/workspace", "containerDiskInGb": '${CONTAINER_SIZE}', "interruptible": true, "cloudType": "COMMUNITY", "env": { "JUPYTER_PASSWORD": "runpod123", "PYTHONPATH": "/workspace", "CUSTOM_VAR": "custom_value" }, "ports": [ "8888/http", "22/tcp" ], "templateId": "'${TEMPLATE_ID}'" }' =================== Prompt examples used: API_URL="https://2kqahvf6g48jji-5001.proxy.runpod.net" echo echo "=== Prompt 1 ===" echo curl -X POST "${API_URL}/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer dummy" -d '{ "model": "koboldcpp", "messages": [ { "role": "user", "content": "Explain quantum computing in simple terms" } ], "max_tokens": 150, "temperature": 0.7, "stream": false }' echo echo "=== Prompt 2 ===" echo curl -X POST "${API_URL}/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer dummy" -d '{ "model": "koboldcpp", "messages": [ { "role": "user", "content": "The future of artificial intelligence is" } ], "max_tokens": 150, "temperature": 0.7, "stream": false }' echo echo "=== Prompt 3 ===" echo curl -X POST "${API_URL}/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer dummy" -d '{ "model": "koboldcpp", "messages": [ { "role": "user", "content": "Write a short story about a robot" } ], "max_tokens": 150, "temperature": 0.7, "stream": false }' Tue, 21 Jul 2026 20:40:56 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://peer.madiator.cloud https://peer.madiator.cloud/lazy-static/avatars/e2927aea-9e0a-454c-9327-68ec87b4da11.png https://peer.madiator.cloud/videos/watch/67731369-4b43-476c-a14d-f15edf8a93b7 All rights reserved, unless otherwise specified in the terms specified at https://peer.madiator.cloud/about and potential licenses granted by each content's rightholder.