Get API key

The open market for production inference.

Providers compete on price and speed to give you the best prices on the market.

Models
85
Markup on provider prices
0%

Offers

Discounts, free models and this week's price drops, on one key.

All 16 discounted models

New models

Added in the last two weeks.

All new models

Latest in the market

Market log
  • New list price

    Gemma 4 26B A4BPrimeIntellect (bf16) list price
    $0.10 in$0.30 out256K ctxvision
  • New list price

    Kimi K3Wafer (fast) list price
    $0.50 in$22.00 out1M ctxvision
  • List price rise

    DeepSeek V4 Flash 0731Alibaba list price
    ↑ 100%
    $0.18 to $0.35 in$0.53 to $1.06 out
  • List price rise

    DeepSeek V4 Flash 0731Alibaba (us-east-1) list price
    ↑ 100%
    $0.19 to $0.37 in$0.56 to $1.12 out
  • List price rise

    DeepSeek V4 Flash 0731Relace (fp4) list price
    ↑ 13%
    $0.042 to $0.048 in$0.17 to $0.16 out
  • List price rise

    DeepSeek V4 Pro 0813Alibaba list price
    ↑ 93%
    $0.58 to $1.12 in$1.74 to $3.37 out
  • List price rise

    DeepSeek V4 Pro 0813Alibaba (us-east-1) list price
    ↑ 100%
    $0.66 to $1.32 in$1.98 to $3.96 out
  • List price rise

    DeepSeek V4 Pro 0813DeepSeek list price
    ↑ 100%
    $0.66 to $1.32 in$1.98 to $3.96 out
  • List price rise

    DeepSeek V4 Pro 0813Relace (fp4) list price
    ↑ 93%
    $0.12 to $0.23 in$8.00 out
  • List price rise

    DeepSeek V4.1 FlashAlibaba list price
    ↑ 100%
    $0.17 to $0.33 in$0.66 to $1.32 out

Anyone can serve

List your endpoint at your own price. It goes live once our automated check passes.

Automatic failover

If a provider errors or times out, the next provider for the same model answers.

Price and speed

Each request goes to the provider that is cheapest and fastest right now.

Data policies

Keep prompts in the EU, or away from providers that train on them or keep them.

Two sides of one market

Callers buy inference at the provider's price. Providers sell it at a price they set.

Buy inference

One key for every model

Call any model through one API and pay the price of the provider that answered.

Kimi K36 providers, $2.565 to $3.45
  1. llm.buzzLowest$2.565
  2. DeepInfra$2.85
  3. Moonshot AI$3.00
  4. Novita$3.00

Input price per million tokens, live

Models
85
Providers
11
Lowest input, 1M
$0.045
  • The OpenAI and Anthropic SDKs work as they are.
  • If a provider fails, the next one for the same model answers.
  • A 5.5% fee when you buy credit, never on tokens.
Sell inference

Your endpoint, your price

List the models you serve at your own price. Requests reach you when you are the best offer.

Serving now+2 more
Providers live
11
In the EU
3
Of your price you keep
100%
  • A model goes live once the automated check passes.
  • You also answer when the provider ahead of you fails.
  • Statements and payouts are in the console.
3providers serve from the EUName them in a request and it stays with them, failover included. On every plan, no upgrade.EU routing

Works with your agents

Point them at api.llm.buzz/api/v1. The console shows what each one spends.

See all apps and agents

How it works

  1. Sign up

    Create an account to get started.

  2. Buy credits

    Credits work with every model and every provider.

  3. Get your API key

    Create a key and point your SDK at llm.buzz.

From the blog

All posts