Verified, then logged.
Sun 22 Nov 2026 · 00:34 KST / 15:34 UTC한국어 | ENSearch /

Breaking02 / IT

Cline v0.0.48 fixes token bug behind rejected GPT-6 requests

Default models also change for several providers, but the release notes do not say whether existing settings are affected.

Over a developer's shoulder, a monitor running a coding tool shows a list of requests marked with red bars, ending in a single green bar. AI-generated image

Cline released Desktop v0.0.48 on 17 Oct 2026 and said it had fixed an error that caused GPT-6 requests to be rejected. The fix affects users of OpenAI and Azure OpenAI-compatible endpoints.

According to the release notes, models such as gpt-6-sol and gpt-6.1-sol now work on those endpoints. Earlier versions of Cline sent max_tokens instead of max_completion_tokens. Every request was rejected as a result.

Compatibility. The release notes do not flag any breaking changes. The default model for the Cline provider has, however, returned to Claude Sonnet 5.5. That model is also back at the top of Cline's recommended list.

The release notes also list a model catalogue refresh that adds the CommandCode and Reflection providers. The default model for 302.AI is now Gemini 4 Argon, and Scaleway's is DeepSeek V4.1 Flash. CoreWeave now defaults to Reflection Beam 1.0.

NanoGPT's default model has also changed, but the release notes do not give the new model's name or say whether existing settings change.