From fe33568ee03b8b43f7f9317dfdf27f051002f871 Mon Sep 17 00:00:00 2001 From: "Dhanji R. Prasanna" Date: Wed, 28 Jan 2026 13:58:14 +1100 Subject: [PATCH] Fix embedded provider max_tokens default (2048 -> 8192) The resolve_max_tokens() function was returning 2048 for embedded providers, which caused responses to be truncated prematurely. Increased to 8192 to allow the provider's own effective_max_tokens() calculation to work properly. --- crates/g3-core/src/provider_config.rs | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/crates/g3-core/src/provider_config.rs b/crates/g3-core/src/provider_config.rs index 4ee4550..d990dc7 100644 --- a/crates/g3-core/src/provider_config.rs +++ b/crates/g3-core/src/provider_config.rs @@ -72,7 +72,7 @@ pub fn resolve_max_tokens(config: &Config, provider_name: &str) -> u32 { "anthropic" => 32000, // Anthropic provider defaults to 32768, we use 32000 "databricks" => 32000, // Databricks is passthru to Anthropic, match its defaults "openai" => 32000, // OpenAI models support large outputs - "embedded" => 2048, // Embedded provider defaults to 2048 + "embedded" => 8192, // Embedded provider: let provider's effective_max_tokens() handle it _ => 16000, // Generic fallback }; let base = get_max_tokens(config, provider_name).unwrap_or(provider_default);