{
  "listings": [
    {
      "slug": "baseten",
      "name": "Baseten",
      "company": "Baseten Labs, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "min_replica 0 scales to zero and a sync request is parked until a replica is ready (1200 s default, then 500). Replica time bills per minute from start-up, and scale_down_delay defaults to 900 s, so each burst bills 15 idle minutes. No cold-start figure is published.",
      "source": "https://docs.baseten.co/deployment/autoscaling/overview",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/baseten"
    },
    {
      "slug": "beam",
      "name": "Beam",
      "company": "Smartshare, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Endpoints scale to zero and bill only while a container runs; cold start and image pull are free. The default keep_warm_seconds of 180 bills three idle minutes after every call unless lowered, the endpoint budget is capped at 600 s, and no cold-start figure is published.",
      "source": "https://docs.beam.cloud/v2/endpoint/keep-warm",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/beam"
    },
    {
      "slug": "cerebrium",
      "name": "Cerebrium",
      "company": "Cerebrium Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "min_replicas 0 scales to zero, cooldown defaults to 60 s, and cold-start time is not billed while model initialisation is. A request queues until a container is ready. disable_auth defaults to true, so the endpoint is public until changed, and no cold-start figure is published.",
      "source": "https://cerebrium.ai/docs/scaling/scaling-apps",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cerebrium"
    },
    {
      "slug": "cloud-run-gpus",
      "name": "Cloud Run GPUs",
      "company": "Google LLC",
      "grade": "B",
      "fit": "caveat",
      "why": "GPU services scale to zero and the GPU is usable about 5 s after an instance starts, with weights loaded on top from Cloud Storage. Instance-based billing charges the whole instance lifetime, idle included, so a request every 15 minutes billed a user all day (acc-36). One L4 or RTX PRO 6000 per instance.",
      "source": "https://docs.cloud.google.com/run/docs/configuring/services/gpu",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/cloud-run-gpus"
    },
    {
      "slug": "hugging-face-inference-endpoints",
      "name": "Hugging Face Inference Endpoints",
      "company": "Hugging Face, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "Scale to zero defaults to 1 hour idle (settable to 15 minutes) and bills nothing at zero replicas, but the docs say scaling from zero takes minutes, the proxy answers 503 unless X-Scale-Up-Timeout holds the request, and users reported endpoints stuck Initializing for a week (acc-39) and no 503 on wake (acc-40).",
      "source": "https://huggingface.co/docs/inference-endpoints/guides/autoscaling",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/hugging-face-inference-endpoints"
    },
    {
      "slug": "koyeb",
      "name": "Koyeb",
      "company": "Koyeb SAS",
      "grade": "D",
      "fit": "caveat",
      "why": "GPU services scale to zero (public preview) after a 5 minute idle period and wake on the next request with a 1 to 5 s VM cold start, plus model load. Light Sleep snapshots are CPU only, the idle period is only adjustable on Pro and above, and no model-level cold-start figure is published.",
      "source": "https://www.koyeb.com/docs/run-and-scale/scale-to-zero",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/koyeb"
    },
    {
      "slug": "modal",
      "name": "Modal",
      "company": "Modal Labs, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "Functions scale to zero by default with a 60 s scaledown_window and the web request waits through the cold start. Users measured 20 to 30 s for demucs (acc-13) and about a minute for ComfyUI (acc-12); GPU memory snapshots are alpha and one user saw 24 of 41 restores segfault (acc-18). Endpoints are public until a proxy token is added.",
      "source": "https://modal.com/docs/guide/scale",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/modal"
    },
    {
      "slug": "replicate-deploy",
      "name": "Replicate Deployments",
      "company": "Replicate, LLC",
      "grade": "B",
      "fit": "caveat",
      "why": "POST /v1/deployments with hardware, min_instances 0 and max_instances gives a private autoscaling endpoint that scales to zero, but instances bill per second for set-up and idle time, prediction data is deleted after an hour, and no cold-start figure or scale-down window is published.",
      "source": "https://replicate.com/docs/topics/deployments/create-a-deployment",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/replicate-deploy"
    },
    {
      "slug": "runpod",
      "name": "Runpod",
      "company": "Runpod, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Workers scale to zero after an idle timeout (5 s console default) and FlashBoot caches starts, but workers bill during cold start and idle. Users measured 63 s engine-ready for Qwen3-8B (acc-11), 12.7 s to first byte for a TTS template (acc-05) and cold starts up to five minutes (acc-06).",
      "source": "https://docs.runpod.io/serverless/overview",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/runpod"
    },
    {
      "slug": "vast-ai",
      "name": "Vast.ai",
      "company": "Vast.ai Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Serverless endpoints recruit marketplace workers from a template with min_workers, inactivity_timeout and cold_mult parameters and pay only for execution, but each new endpoint needs a benchmarking evaluation period across GPU classes, workers are peer hosts, and no cold-start figure is published.",
      "source": "https://docs.vast.ai/guides/serverless/index",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/vast-ai"
    },
    {
      "slug": "verda",
      "name": "Verda",
      "company": "DataCrunch Oy",
      "grade": "C",
      "fit": "caveat",
      "why": "Serverless Containers scale to zero when idle with queue-based autoscaling and a vLLM quickstart, billing every minute a replica is up including spin-up and spin-down. The scale-down delay must be tuned by hand and no cold-start figure is published.",
      "source": "https://docs.verda.com/containers/overview",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/verda"
    },
    {
      "slug": "coreweave",
      "name": "CoreWeave",
      "company": "CoreWeave, Inc.",
      "grade": "C",
      "fit": "no",
      "why": "Dedicated Inference autoscaling requires min at least 1 and states scale-to-zero is not supported; CKS node pools are whole 8-GPU nodes billed by the hour. Nothing bills zero when idle.",
      "source": "https://docs.coreweave.com/products/inference/scaling.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/coreweave"
    },
    {
      "slug": "crusoe-cloud",
      "name": "Crusoe Cloud",
      "company": "Crusoe Technologies LLC",
      "grade": "B",
      "fit": "no",
      "why": "VMs bill per second while running with no autoscaling, and self-serve deployments reserve GPUs by the hour for Crusoe's catalogue models and LoRA adapters, not a user's container or handler.",
      "source": "https://docs.crusoe.ai/compute/virtual-machines/overview",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/crusoe-cloud"
    },
    {
      "slug": "gpusmarket",
      "name": "GPUsMarket",
      "company": "Tyga.Cloud Ltd",
      "grade": "E",
      "fit": "no",
      "why": "Rentals serve the host's Ollama or OpenAI-compatible models and bill per second until stopped; there is no custom handler or container with scale to zero. Cloud pods can be paused but not autoscaled.",
      "source": "https://gpusmarket.com/llms.txt",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/gpusmarket"
    },
    {
      "slug": "hyperbolic",
      "name": "Hyperbolic",
      "company": "Hyperbolic Labs, Inc.",
      "grade": "E",
      "fit": "no",
      "why": "Hourly VM and bare-metal rentals only; no containers, endpoints, autoscaling or scale to zero. The separate Inference product serves Hyperbolic's hosted models, not the user's.",
      "source": "https://hyperbolic.ai/docs/on-demand/overview",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/hyperbolic"
    },
    {
      "slug": "lambda",
      "name": "Lambda Cloud",
      "company": "Lambda, Inc.",
      "grade": "C",
      "fit": "no",
      "why": "On-demand VMs billed by the minute until terminated; no containers, endpoints, autoscaling or scale to zero.",
      "source": "https://docs.lambda.com/public-cloud/on-demand/",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/lambda"
    },
    {
      "slug": "massed-compute",
      "name": "Massed Compute",
      "company": "Massed Compute, Inc.",
      "grade": "E",
      "fit": "no",
      "why": "Hourly GPU VMs only; no containers, endpoints, autoscaling or scale to zero, and a stopped VM still bills the full hourly rate.",
      "source": "https://vm-docs.massedcompute.com/api/v1",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/massed-compute"
    },
    {
      "slug": "nebius-ai-cloud",
      "name": "Nebius AI Cloud",
      "company": "Nebius B.V.",
      "grade": "BB",
      "fit": "no",
      "why": "A Serverless AI endpoint is one container on one VM that runs until stopped; stopping bills nothing but there is no autoscaling or scale to zero, so an idle endpoint keeps paying GPU rates until a call stops it.",
      "source": "https://docs.nebius.com/serverless/overview",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nebius-ai-cloud"
    },
    {
      "slug": "northflank",
      "name": "Northflank",
      "company": "Northflank Ltd",
      "grade": "C",
      "fit": "no",
      "why": "GPU services autoscale between a minimum and maximum instance count but have no scale to zero; a service bills until paused or deleted, so an idle endpoint keeps paying.",
      "source": "https://northflank.com/docs/v1/application/scale/autoscale-deployments.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/northflank"
    },
    {
      "slug": "thunder-compute",
      "name": "Thunder Compute",
      "company": "Thunder GPU, Inc.",
      "grade": "D",
      "fit": "no",
      "why": "Persistent instances billed per minute while they run and sandboxes with a fixed lifetime; no autoscaling endpoint or scale to zero.",
      "source": "https://www.thundercompute.com/docs/billing",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/thunder-compute"
    },
    {
      "slug": "together-gpu-clusters",
      "name": "Together GPU Clusters",
      "company": "Together Computer, Inc.",
      "grade": "C",
      "fit": "no",
      "why": "Clusters bill hourly from creation until deleted and cannot be paused or scaled to zero; there is no endpoint product here.",
      "source": "https://docs.together.ai/docs/gpu-clusters-billing",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/together-gpu-clusters"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.8",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "useCase": {
    "accounts": 21,
    "after": [
      "inference.tool-loop",
      "speech-to-text.long-recording-batch",
      "voice-agents.low-latency-pipeline-build"
    ],
    "area": "models",
    "asks": [
      "deploy this model as an API I can call",
      "run Whisper or a diffusion model behind an endpoint that costs nothing when idle",
      "host my fine-tune so my app can call it"
    ],
    "before": [
      "gpu-compute.package-and-deploy",
      "open-weights.fine-tune-base"
    ],
    "blockers": [
      "cold start of a minute or more for a request-response app, so the user gives up before the first token",
      "endpoint stuck initialising with no logs and no 503, so the caller waits on nothing",
      "a warm replica or keep-warm window bills while idle, so a request every 15 minutes costs a full day",
      "snapshot restore that fails intermittently and falls back to a full cold start",
      "endpoint public by default until a token is added"
    ],
    "capabilities": [
      "compute.serverless",
      "compute.endpoints",
      "compute.gpu",
      "compute.containers",
      "compute.scale-to-zero"
    ],
    "category": "gpu-compute",
    "categoryName": "GPU \u0026 serverless compute",
    "checked": "2026-11-12",
    "companies": {
      "fit": 0,
      "fitWithLimit": 11,
      "tagged": 24
    },
    "competition": "none",
    "fits": null,
    "fitsWithLimit": [
      "baseten",
      "beam",
      "cerebrium",
      "cloud-run-gpus",
      "hugging-face-inference-endpoints",
      "koyeb",
      "modal",
      "replicate-deploy",
      "runpod",
      "vast-ai",
      "verda"
    ],
    "id": "gpu-compute.serverless-model-endpoint",
    "listings": {
      "caveat": 11,
      "no": 10,
      "notRead": 0,
      "tagged": 24,
      "unchecked": 0,
      "yes": 0
    },
    "name": "Serve my own model as an endpoint that scales to zero",
    "status": "draft",
    "success": "A first request after scale to zero returns a correct answer including weight load, a warm request returns, and the endpoint bills nothing after the idle window with no replica left running. Measured: the first request after scale to zero answers within 60 seconds, and a warm request within 2 seconds plus inference.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/gpu-compute.serverless-model-endpoint.json",
    "weight": 28,
    "weightBasis": "21 of 48 accounts are about an endpoint on serverless GPU: cold start, idle billing, wake from zero. The most written-about job on Hacker News and in the issue trackers read."
  }
}
