Model provider cost

Host Step 3.7 Flash on RunPod

Step 3.7 Flash needs 1x 120GB+ GPU, and RunPod's current cheapest qualifying row is MI300X at $2.39/hr.

1x 120GB+ GPU 201B params Efficient heavy model Apache 2.0

All LLM costs All RunPod GPUs

Cheapest on this provider

$2.39/hr

MI300X

Monthly estimate

$1,745/mo

730 hours at the current median

VRAM baseline

120GB

1x 120GB+ GPU

Qualifying rows

Updated Jul 27, 2026

Live hosting options

RunPod rows that can host Step 3.7 Flash

The cheapest tracked way to host Step 3.7 Flash on RunPod is MI300X at $2.39/hr. The overall tracked market floor is $2.29/hr on Lambda, so RunPod is $0.10/hr above the current floor.

Price source RunPod GraphQL API Provider API pricing; availability and spot/community rows can move quickly.

Methodology Freshness

GPU	VRAM	Per GPU	Estimated hourly	Estimated monthly	Offers	Updated
MI300X	192GB	$2.39/hr	$2.39/hr	$1,745/mo	1	Jul 27, 2026 RunPod GraphQL API 1 offer · Jul 27, 2026 global aggregate
H200 NVL	141GB	$3.79/hr	$3.79/hr	$2,767/mo	1	Jul 27, 2026 RunPod GraphQL API 1 offer · Jul 27, 2026 global aggregate
H200	141GB	$4.39/hr	$4.39/hr	$3,205/mo	1	Jul 27, 2026 RunPod GraphQL API 1 offer · Jul 27, 2026 global aggregate
B200	192GB	$5.89/hr	$5.89/hr	$4,300/mo	1	Jul 27, 2026 RunPod GraphQL API 1 offer · Jul 27, 2026 global aggregate

Model fit

Why this setup does or does not fit

VRAM floor

1x 120GB+ GPU

1x 141GB H200 or 2x 80GB GPUs with a fast interconnect. Long prompts, batching, and KV cache can require extra headroom.

Model quality

Efficient heavy model

Interesting when multimodal agent throughput matters enough to justify a much larger resident weight footprint.

Operational note

StepFun 201B

The official card lists a 120GB minimum for Q4 unified-memory deployment; sparse activation does not remove that memory floor.

FAQ

Step 3.7 Flash on RunPod FAQ

Can I host Step 3.7 Flash on RunPod?

The cheapest tracked way to host Step 3.7 Flash on RunPod is MI300X at $2.39/hr.

What GPU memory does Step 3.7 Flash need?

Our baseline for Step 3.7 Flash is 1x 120GB+ GPU. The practical recommendation is 1x 141GB H200 or 2x 80GB GPUs with a fast interconnect.

Is RunPod the cheapest provider for Step 3.7 Flash?

The overall tracked market floor is $2.29/hr on Lambda, so RunPod is $0.10/hr above the current floor.

How fresh is this RunPod Step 3.7 Flash cost page?

This page recalculates from the latest tracked on-demand rows. The freshest qualifying RunPod row shown here is from Jul 27, 2026.

Next research paths

Compare this setup

Agent-assisted setup

Hand this runbook to Claude Code or Codex

Open a terminal in the repository where you want the deployment files, start claude or codex, then paste this prompt. It asks the agent to verify sources and stop before it creates billable infrastructure.

Claude Code quickstart ↗ Codex docs ↗

Deployment prompt

Use the infrastructure or model context on this page to create a reproducible open-model deployment. Use this guide as the starting context: https://www.getflops.ai/llms/step-3.7-flash/runpod. Read the linked model card and provider documentation before choosing hardware or runtime settings. Open every linked primary source and flag any mismatch instead of guessing. Create a deployment folder containing README.md, .env.example with no secrets, a pinned start script or infrastructure manifest, and smoke-test.sh. Make the endpoint OpenAI-compatible where the runtime supports it. Run local/static validation, estimate the billable resources, and stop before provisioning paid infrastructure until I approve.

Guardrails included No secrets in files · verify primary docs · approval before spend