Evaluation · free, self-serve
Prove it on your own models.
Run the scored Alphabet Test on the approved languages, track every result, re-test to watch your model improve, and compare models head-to-head. You get the score and actionable feedback — never the underlying inventory.
- ✓ Self-serve — confirm your email and you're in
- ✓ Run by web, API, or MCP — your agent can run it headless
- ✓ Stored results + improvement tracking + team sharing
- ✓ Rates & coaching only — no licensed data ever
Bemba + Chewa today, plus 5 languages on L26 Lite
Scored on the L26 benchmark
Renewable — no expiry