Forumics logo Forumics
Issue 01

Back to issue

All Cline known issues

CLAIM C168 Supported 2026-11-20

GitHub issue #15778 in cline/cline, opened 2026-11-18 by sarvesh-pathak, records a report that the built-in OpenAI Compatible provider points at api.openai.com v1 and defaults to gpt-4o, while the OpenAI deprecations page lists GPT-4o with a shutdown date of Jan 14, 2027 and gpt-6-luna as the replacement, so anyone who accepts the default with an OpenAI key gets a model-not-found error after that date; the report adds that changing the ID alone will not be enough, since gpt-6-luna Chat Completions supports function calling only with reasoning_effort set to none, and issue 12866 describes the same 400 on gpt-5.6-luna and is still open; the report notes it has not run Cline against gpt-6-luna and that the shutdown date is in the future.

Verbatim Excerpts from Cited Records
“The built-in "OpenAI Compatible" provider points at `https://api.openai.com/v1` and defaults to `gpt-4o`:”
— sarvesh-pathak · cline/cline (issue 15778 opening body, What happened section) · 2026-11-18T18:57:31Z ·GitHub Issue ↗
“OpenAI's deprecations page lists GPT-4o (`gpt-4o-2024-08-06`, `gpt-4o-2024-11-20`) with a shutdown date of Jan 14, 2027 and `gpt-6-luna` as the replacement (https://developers.openai.com/api/docs/deprecations, read 18 Nov 2026). After that date, anyone who accepts the default model with an OpenAI key gets a model-not-found error.”
— sarvesh-pathak · cline/cline (issue 15778 opening body, What happened section) · 2026-11-18T18:57:31Z ·GitHub Issue ↗
“**Changing the ID alone won't be enough.** The `gpt-6-luna` model page says "Chat Completions supports function calling only with `reasoning_effort` set to `none`" (https://developers.openai.com/api/docs/models/gpt-6-luna, read 18 Nov 2026). This provider uses Chat Completions with tools. #12866 describes the same 400 on `gpt-5.6-luna` ("Function tools with reasoning_effort are not supported ... set reasoning_effort to 'none'") and is still open. So moving the default to `gpt-6-luna` probably needs #12866, or an explicit `reasoning_effort: "none"` for this model, to land with it. I haven't run Cline against `gpt-6-luna`.”
— sarvesh-pathak · cline/cline (issue 15778 opening body, What happened section) · 2026-11-18T18:57:31Z ·GitHub Issue ↗
“The CLI onboarding pre-fills that default when you pick the provider (`apps/cli/src/tui/views/onboarding/controller.ts` L368, L375, L390). It is also the model #15008 saw being picked when no model ID was saved.”
— sarvesh-pathak · cline/cline (issue 15778 opening body, What happened section) · 2026-11-18T18:57:31Z ·GitHub Issue ↗
Counter-record: Issue open as of the fetch; 2 comments were read. linear-code[bot] linked the Linear issue CLINE-3807. Shiv-61 on 2026-11-19 agreed that bumping defaultModelId to gpt-6-luna or gpt-5.6-luna will immediately hit the 400 error on tool calls unless reasoning_effort none is explicitly sent, offered to help with a PR, and asked maintainers about the preferred replacement default, whether the fix from issue 12866 should be revived into the same PR or land first, and whether self-hosted setups should keep pre-filling an OpenAI-specific default. No maintainer response appears in the surveyed thread. The claim documents only the original report; the shutdown date is in the future and not yet observed. Falsification probe: If GitHub issue #15778 is deleted without archive, if the author edits the opening body so the quoted sentences no longer match verbatim, or if the author posts a retraction of the report, this claim drops to Withdrawn. Audit verification: Verified 100% continuous verbatim substring against official GitHub REST API by scripts/verify_claim_verbatim.py.