What this shortlist means
The inclusion rules are visible and intentionally narrow.
- 01
Inclusion rule
The model carries the reviewed Code capability tag.
- 02
Inclusion rule
The catalog identifies the API protocol as OpenAI compatible.
- 03
Inclusion rule
Candidates are ranked by the current editorial priority, not sponsorship.
Important limit: OpenAI-compatible chat completions do not prove support for every Codex feature or the Responses API. Verify the exact endpoint and tool behavior before relying on a route.
Matching candidates
Live catalog results with provider evidence dates and the limits you should read before signup.
| Model and provider | Why it matches | Evidence and limits | Next steps |
|---|---|---|---|
| GLM-5.2 NVIDIA NIM z-ai/glm-5.2 | Code tag + OpenAI-compatible protocol Developer access | Official sources: 2026-07-28 Not listed in the official quickstart Up to 40 RPM; account limits may differ Independent Canary pending | |
| Qwen3.6 27B GroqCloud qwen/qwen3.6-27b | Code tag + OpenAI-compatible protocol No-card free plan | Official sources: 2026-07-28 Not required for the Free plan Model and organization specific Independent Canary pending | |
| DeepSeek R1 SiliconFlow deepseek-ai/DeepSeek-R1 | Code tag + OpenAI-compatible protocol Starter credits / free-model policy | Official sources: 2026-07-28 Not listed as required for starter access Model and account specific Independent Canary pending | |
| GPT-OSS 120B Cerebras Inference gpt-oss-120b | Code tag + OpenAI-compatible protocol Payment-verified starter credits | Official sources: 2026-07-28 Verified payment method required for starter credits Plan and model specific Independent Canary pending |