Skip to main content
Surplus Intelligence
Marketplace

The inference exchange

Open market
for AI models

Your AI calls are routed to the best price.Your idle capacity is sold to the highest bidder.

Anthropic

Direct price

$25.00

Per 1M output tokens

With Surplus

$0.4020

Per 1M output tokens

Surplus saves you $24.60 with a 98.4% discount.
Get API key
Build with agents

Access leading models

Anthropic
OpenAI
Google
Meta
xAI
DeepSeek
Mistral AI
Alibaba
Qwen
NVIDIA
Moonshot
MiniMax
Z.ai
ByteDance
Black Forest Labs
Kuaishou
Runway
ElevenLabs
Luma AI
PixVerse
Lightricks
Tencent
Vidu
Arcee AI
LongCat
Venice AI
Uncensored AI
Aion Labs
Sakana
Fal AI
Fireworks
OpenRouter
Topaz Labs
TypeSafe

Market analytics

The market, in motion

Explore analytics

Discounted market liquidity

Live

$24B+

Listed across 22,047 offers and 430 models

Largest marketLong tail
Listed liquidity across 358 active markets, highest to lowest · largest $1B · median $345 · smallest $0.25

Token volume · 28 days

3.5T↑ 31.4%

input tokens

62.1B↑ 31.9%

output tokens

2.8T cache tokens

InputOutputCumulative progress

Requests routed · 28 days

67.9M↑ 27.7%

3.6M in the last 24 hours

28 days agoToday

Realized savings · 7 days

89.0%

Mean discount · $28.5K saved in 24h↓ 14.1%

Peak: 95-100%

Two-minute setup

Integrate in 2 minutes

Change your base URL.
Cut your inference costs.

Build with agents

Built for coding agents. llms.txt explains the endpoints, authentication, models, and payments.

Works with your SDK. Use OpenAI-compatible clients or Anthropic Messages.

Keeps keys private. Your API key stays in an environment variable.

quickstart.py
1from openai import OpenAI
2import os
3 
4client = OpenAI(
5 # base_url="https://openrouter.ai/api/v1",
6 base_url="https://api.surplusintelligence.ai/v1",
7 api_key=os.environ["SURPLUS_API_KEY"],
8)

Ask the community

Trusted by real builders & real products

Developers building with Surplus have cut their inference costs by 83%.You could, too.

App

Surplus Optimizer

by michielpostnl

Agent

AgentOS

by useAgentOS

App

Smarter Intel

by Daemon

App

Claw Hunter

by clawhuntersol

App

CELPIP Practice

by cankeremgurel

Tool

bu0y

by Crypt0Gnome

App

tokensto.cash

by usdctofiat

App

Otto

by Otto

App

Ruvicode

by asvm

App

surp.ivc.lol

by ivcained

App

Bankr LLM Gateway

by Bankr

App

Sleuth AI

by 99barzzz

Tool

OpenClaw Smart Router

by mykclawd

App

Harness

by tryharness

Agent

FrostyFi x402Fetch

by frosty_onchain

Tool

Pi Surplus Intelligence

by gkgoat1

Tool

Surplus AI SDK Provider

by dylsteck

App

Surplus Bench

by euwars

Calculator

See how much you can save

Pick a model. Move the slider. See the difference.

Monthly usage
$1,000.00per month
$100$10,000

Inference bill

Surplus Intelligence

Monthly estimate

Direct vs. Surplus

Model
Anthropic
Claude Opus 4.6
Usage

$1,000.00 / month

Your current provider

Provider list price

$1,000.00

Discount offered via Surplus

@ active healthy market price

−$983.92

Total with Surplus

Estimated monthly bill

$16.08

98.4% cut

surplusintelligence.ai98.4% off

How it works

Built for both sides of the market

Anthropic

Claude Opus 4.6

$25.00$0.4020/ 1M output

−98.4%
DeepSeek

DeepSeek V4.1 Flash

$1.20$0.0018/ 1M output

−99.9%
DeepSeek

DeepSeek V4 Pro

$1.91$0.0043/ 1M output

−99.8%
Google

Gemini 3.8 Flash

$3.75$0.0121/ 1M output

−99.7%
Google

Gemma 4 26B A4B IT

$0.30$0.0013/ 1M output

−99.6%
Explore all discounts on Model Marketplace

Get your Surplus key

One key gives you access to every model.

Integrate in minutes

Point your existing SDK at one compatible base URL.

Route and save

Each call finds the healthiest offer, at your minimum discount

Get your API keyExplore live prices

The market is open

Every model. One market. Your terms.

Buy inference at the best available price or turn unused compute into revenue.

Explore the marketRead the docs