Add NVIDIA provider, and improve declarative provider UX (#8798)

Signed-off-by: jh-block <jhugo@block.xyz>
2026-04-24 15:48:30 +02:00
parent 8f16ec61e4
commit 4065d44ef7
9 changed files with 154 additions and 21 deletions
@@ -158,7 +158,11 @@ impl ModelConfig {
                self.context_limit = Some(canonical.limit.context);
            }
            if self.max_tokens.is_none() {
-                self.max_tokens = canonical.limit.output.map(|o| o as i32);
+                self.max_tokens = canonical
+                    .limit
+                    .output
+                    .filter(|&output| output < canonical.limit.context)
+                    .map(|output| output as i32);
            }
            if self.reasoning.is_none() {
                self.reasoning = canonical.reasoning;
@@ -491,6 +495,20 @@ mod tests {
            assert_eq!(config.max_tokens, Some(1_000));
        }

+        #[test]
+        fn skips_canonical_output_limit_when_it_equals_context_limit() {
+            let _guard = env_lock::lock_env([
+                ("GOOSE_MAX_TOKENS", None::<&str>),
+                ("GOOSE_CONTEXT_LIMIT", None::<&str>),
+            ]);
+            let config =
+                ModelConfig::new_or_fail("moonshotai/kimi-k2.5").with_canonical_limits("nvidia");
+
+            assert_eq!(config.context_limit, Some(262_144));
+            assert_eq!(config.max_tokens, None);
+            assert_eq!(config.max_output_tokens(), 4_096);
+        }
+
        #[test]
        fn unknown_model_leaves_fields_none() {
            let _guard = env_lock::lock_env([