Made with Kleap
Mark Get started
Mark v4 β€” now generally available

Intelligence that leaves a mark.

Mark is the AI platform built for builders. Frontier reasoning models, lightning-fast inference, and a developer experience that finally feels like the future.

mark.terminal β€” reasoning mode
// prompt
"Design a payment system for 10M users"
β†’ reasoning trace...
Analyzing throughput targets...
Selecting shard strategy...
Generating architecture βœ“
Response 847ms
Horizontal sharding by tenant ID
Eventual consistency with idempotency keys
Idempotent webhooks & retry budget
Saga pattern for cross-shard transfers

Trusted by teams shipping faster

NOVA Helios STRATA.IO AXIOM/ LUMEN β—’ Vertex Quanta

Capabilities

A single API for every model you'll ever need.

Frontier reasoning

Multi-step planning, chain-of-thought, and tool use that holds up under pressure. Mark reasons for minutes, not seconds β€” and gets the hard problems right.

96.4%
SWE-bench verified
1.2M
context tokens

Sub-200ms inference

Streaming first, globally distributed, edge-cached. Mark responds before your users notice they asked.

Vision, voice, code

Native multimodal β€” drop in images, audio, PDFs, repositories. Mark understands them all in one context window.

Models

Pick the right brain for the job.

Four specialized models, one unified API. Route by complexity, latency, or cost β€” Mark handles the rest.

Compare all models
M
Mark-4 Ultra
Deep reasoning for complex, multi-step problems
M
Mark-4
Production workhorse β€” best balance of capability and speed
M
Mark-Mini
Lightning fast for classification, extraction, chat
M
Mark-Vision
Native multimodal understanding of images and video
Abstract neural network

Built different

Trained on the most diverse compute cluster on Earth.

Mark runs on a custom-built inference fabric spanning 14 regions.

Pricing

Simple, by token.

No seats. No surprises. Pay for what you use, scale when you're ready.

Free

For prototypes & side projects

$0 /mo
  • 1M free tokens / month
  • All 4 models
  • Community support
  • Rate limit: 60 req/min
Start free
Most popular

Pro

For shipping teams

$49 /mo + usage
  • 50M tokens included
  • Priority inference queue
  • Email support, 24h SLA
  • Rate limit: 10k req/min
  • Fine-tuning credits
Start 14-day trial

Enterprise

For scale & compliance

Custom
  • Volume pricing
  • Dedicated capacity
  • SOC 2, HIPAA, custom DPA
  • Dedicated solutions engineer
  • On-prem deployment option
Talk to sales

Ready to leave your mark?

Join 40,000+ teams building the next generation of intelligent software with Mark.

Contact

Let's build something impossible.

Tell us what you're working on. We respond within one business day, usually faster.

Email
hello@mark.ai
HQ
San Francisco Β· Berlin Β· Tokyo