From 56102ff6422ceafdeb349ba12bb08691660471bb Mon Sep 17 00:00:00 2001 From: Johannes Loher Date: Wed, 18 Mar 2026 05:52:16 +0100 Subject: [PATCH] fix(core): detect vLLM context overflow errors (#17763) Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com> --- packages/opencode/src/provider/error.ts | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/packages/opencode/src/provider/error.ts b/packages/opencode/src/provider/error.ts index dd2554489..7a171f4db 100644 --- a/packages/opencode/src/provider/error.ts +++ b/packages/opencode/src/provider/error.ts @@ -13,7 +13,7 @@ export namespace ProviderError { /input token count.*exceeds the maximum/i, // Google (Gemini) /maximum prompt length is \d+/i, // xAI (Grok) /reduce the length of the messages/i, // Groq - /maximum context length is \d+ tokens/i, // OpenRouter, DeepSeek + /maximum context length is \d+ tokens/i, // OpenRouter, DeepSeek, vLLM /exceeds the limit of \d+/i, // GitHub Copilot /exceeds the available context size/i, // llama.cpp server /greater than the context length/i, // LM Studio @@ -21,6 +21,8 @@ export namespace ProviderError { /exceeded model token limit/i, // Kimi For Coding, Moonshot /context[_ ]length[_ ]exceeded/i, // Generic fallback /request entity too large/i, // HTTP 413 + /context length is only \d+ tokens/i, // vLLM + /input length.*exceeds.*context length/i, // vLLM ] function isOpenAiErrorRetryable(e: APICallError) {