diff --git a/assets/tokyo.yaml b/assets/tokyo.yaml
index 3970dbc..8ce5e07 100644
--- a/assets/tokyo.yaml
+++ b/assets/tokyo.yaml
@@ -26,6 +26,33 @@
   prohibited_uses: ''
   monitoring: unknown
   feedback: none
+- type: model
+  name: Reflection Llama-3.1 70B
+  organization: Unknown
+  description: Reflection Llama-3.1 70B is an open-source LLM, trained with a new technique called Reflection-Tuning that teaches a LLM to detect mistakes in its reasoning and correct course. The model was trained on synthetic data generated by Glaive.
+  created_date: 2024-09-28
+  url: https://huggingface.co/mattshumer/Reflection-70B
+  model_card: https://huggingface.co/mattshumer/Reflection-70B
+  modality:
+    explanation: you can sample from Reflection Llama-3.1 70B using the same code, pipelines, etc. as any other Llama model. This implies that the model takes text input and produces text output.
+    value: text; text
+  analysis: Unknown
+  size:
+    explanation: The size is in the model's name, Reflection Llama-3.1 70B.
+    value: 70B parameters
+  dependencies: [Glaive, Llama 3.1]
+  training_emissions: Unknown
+  training_time: Unknown
+  training_hardware: Unknown
+  quality_control: The model uses a Reflection-Tuning technique which allows it to notice mistakes in its reasoning and correct them.
+  access:
+    explanation: "Reflection Llama-3.1 70B is an open-source LLM."
+    value: open
+  license: Llama 3.1 Community License Agreement
+  intended_uses: The model is intended for complex reasoning and reflection tasks. It is designed to separate its internal thoughts and reasoning from its final answer.
+  prohibited_uses: Unknown
+  monitoring: Unknown
+  feedback: Unknown
 - type: model
   name: Pixtral 12B
   organization: Unknown