text model
GPT-5.6
OpenAIThe model behind HumanLike Coworker — three tiers, one API surface.
GPT-5.6 is what actually runs inside HumanLike Coworker today. We expose it as three named tiers rather than raw effort settings, because in practice teams pick a speed/quality point once and keep it: Luna for everyday actions, Terra for drafting and tool use, Sol when the answer has to be right. Luna is the default for new workspaces.
Variants
3 ways to run it.
Same family, different quality/cost rung. Picking the wrong one is the most common way teams overpay.
Sol
Highgpt-5.6-solHighest-capability GPT-5.6 for demanding work.
Complex analysis · Final scoring passes
Terra
Balancedgpt-5.6-terraBalanced GPT-5.6 intelligence for drafting and tool use.
Drafting · Multi-tool agent steps
Luna
Fastgpt-5.6-lunaFast, efficient GPT-5.6 for everyday Coworker actions. The default.
Live chat · Quick rewrites · Routing
Fit
Two audiences, one model.
The same model is a different proposition depending on what you point it at.
Recruitment fit
The pragmatic hiring default — Terra handles almost every screening and drafting task at a fraction of frontier cost.
Best for
- CV screening at volume
- Candidate Q&A
- Interview note summarisation
Strengths
- Three fixed tiers make cost per hire easy to model
- Already wired into HumanLike, so nothing to integrate
- Luna is fast enough to sit behind a live candidate chat
Watch out for
- Sol is not a substitute for a human on final decisions
- Public benchmark coverage for this alias is thin
The numbers
Measured, not asserted.
HumanLike evals
Run on our own harness.
- Default Coworker chat model
- LunaOur eval
Luna is the shipped default for new Coworker workspaces.
Published benchmarks
Vendor and independent figures.
- MMLU-Pro
- TBDEstimate
