crof.ai

Every frontier model. One API.

Access the best open-source LLMs through an OpenAI-compatible endpoint. Pay per token — no subscriptions, no commitments, no markup.

greg-2-ultra
from $3.00 / M tokens
greg-2-super
from $1.50 / M tokens
glm-5.2
from $0.25 / M tokens
// why crof.ai

Simple by design

One endpoint. Every model. Transparent pricing. Ship in minutes, not weeks.

OpenAI-compatible

Swap the base URL. Keep your existing SDK. Zero migration cost — most integrations work by changing one line.

Pay per token

No subscriptions. No commitments. No minimums. Only pay for what you use, at the fairest rates available.

$

Lowest prices

We regularly have the lowest prices in the industry. Open-source models at cost — frontier models at a fraction of what the big labs charge.

Low latency

Multi-provider routing picks the fastest path for every request. Built-in failover keeps your app running.

// the lineup

Featured models

Our own Greg family, plus the best open-source models from every major lab. One API covers them all.

Our model
greg-2-ultra
CrofAI · frontier
$3.00/ M input tokens
Input$3.00 / M
Output$10.00 / M
229K context · reasoning · UI design · tools
Start using
Our model
greg-2-super
CrofAI · best balance
$1.50/ M input tokens
Input$1.50 / M
Output$5.00 / M
229K context · reasoning · UI design · tools
Start using
glm-5.2
Zhipu AI · open-source
$0.25/ M input tokens
Input$0.25 / M
Output$2.20 / M
128K context · reasoning · tools
Start using
Every model: OpenAI-compatible API pay per token built-in failover
// get started

Minutes to your first call

Stop wrestling with complex SDKs. Copy, paste, ship.

  • 1Get an API key — free account, no credit card required
  • 2Set the base URL — point your OpenAI SDK at crof.ai
  • 3Pick a model — from the list or let us route for you
  • 4Ship it — that's it. Really.
quickstart.sh
# your first CrofAI call
curl https://ai.nahcrof.com/v1/chat/completions \
  -H "Authorization: Bearer $CROF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "greg-2-super",
    "messages": [
      { "role": "user", "content": "say hello" }
    ]
  }'
// pricing

Pay per token. Nothing more.

No subscriptions. No hidden fees. No markup. Just the model's cost, passed through at the lowest rate we can offer.

Model $ / M in $ / M out Speed
glm-5.2 Latest
Q8_0 · 1M context
$0.25 /M
$2.20 /M
~50 t/s
kimi-k2.7-code
int4 · 262K context
$0.55 /M
$2.25 /M
~100 t/s
greg-2-ultra Our model
int4 · 229K context
$3.00 /M
$10.00 /M
~40 t/s

Start building today.

Free account. First call in minutes. No credit card required.