Cheap enough to use freely

Built for an agent’s chores. Set it as small_model in opencode and titles and compaction become nearly free. It struggles alone on complex work — it is not meant to be the main driver.

Context window

500

turns / 1,000 crystals (estimate)

1×

Cost vs the cheapest model

Good for

Subagents and context gathering
Titles, summaries, commit messages
High-volume, low-complexity edits

Not for

Multi-file refactors
Long runs of dozens of steps

Rates

Published price per million tokens — what your crystals are charged at. It is the upstream official list price; check it yourself.

Cache read$0.0028$0.0028

Input$0.1400$0.1400

Output$0.2800$0.2800

Turns per the same amount of crystals.

fast

pro

max

Turn counts are derived from the published rates, assuming a typical agent request of 30k input and 2k output.

How to use it. Put it in an environment variable to call it by name, bypassing the alias default.

Environment variables

export ANTHROPIC_BASE_URL="https://api.pixcode.ai/anthropic"
export ANTHROPIC_AUTH_TOKEN="sk-px-..."
export ANTHROPIC_MODEL="pixcode-fast"

Other tools are covered in the docs.

Which plans include it.

Starter $19/moPro $49/moMax $129/mo

Start with Starter See all models