diff --git a/providers/baseten/models/zai-org/GLM-5.3-Fast.toml b/providers/baseten/models/zai-org/GLM-5.3-Fast.toml
new file mode 100644
index 00000000000..bbee46d8d36
--- /dev/null
+++ b/providers/baseten/models/zai-org/GLM-5.3-Fast.toml
@@ -0,0 +1,20 @@
+# Effort: top-level reasoning_effort = none|low|high|max (default high).
+# none requests a direct answer but may still return short reasoning_content;
+# chat_template_args.enable_thinking=false is rejected, so no separate toggle.
+# https://docs.baseten.co/inference/model-apis/reasoning
+base_model = "zhipuai/glm-5.3"
+name = "GLM 5.3 Fast"
+description = "GLM vision model for visual reasoning, documents, and multimodal agents"
+attachment = true
+reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }]
+
+[cost]
+input = 2.1
+output = 6.6
+
+[limit]
+context = 1_048_576
+output = 262_144
+
+[modalities]
+input = ["text", "image"]