Model comparison
satp-policy-v4.27 vs GPT-5
Capability radar, value bars, and side-by-side posture.
satp-policy-v4.27
Not scored
value score
GPT-5
94
value score
satp-policy-v4.27 ctx
—
GPT-5 ctx
400K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- GPT-594
Attribute tape
| Field | satp-policy-v4.27 | GPT-5 |
|---|---|---|
| Developer | ChristianZ97 | OpenAI |
| Context | — | 400K |
| Modalities | Text, Code, Image, Audio | |
| Openness | — | Proprietary Model |
| Speed | — | Fast |
| Price | — | Mid-High |
| Value score | — | 94 |
| Summary | A reinforcement-learning policy checkpoint for Lean-style automated theorem proving, part of the SATP/Aesop-RL model family developed by ChristianZ97. The model selects proof tactics and lemma configurations during formal proof search. | OpenAI frontier multimodal model. |