diff --git a/README.md b/README.md
index 0bb0bf8..b5287fd 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,6 @@
 # CLIP
 
-[[Blog]](https://openai.com/blog/clip/) [[Paper]](https://cdn.openai.com/papers/Learning_Transferable_Visual_Models_From_Natural_Language_Supervision.pdf) [[Model Card]](model-card.md) [[Colab]](https://colab.research.google.com/github/openai/clip/blob/master/notebooks/Interacting_with_CLIP.ipynb)
+[[Blog]](https://openai.com/blog/clip/) [[Paper]](https://arxiv.org/abs/2103.00020) [[Model Card]](model-card.md) [[Colab]](https://colab.research.google.com/github/openai/clip/blob/master/notebooks/Interacting_with_CLIP.ipynb)
 
 CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pairs. It can be instructed in natural language to predict the most relevant text snippet, given an image, without directly optimizing for the task, similarly to the zero-shot capabilities of GPT-2 and 3. We found CLIP matches the performance of the original ResNet50 on ImageNet “zero-shot” without using any of the original 1.28M labeled examples, overcoming several major challenges in computer vision.
 
diff --git a/clip/clip.py b/clip/clip.py
index 5c21a21..76f241b 100644
--- a/clip/clip.py
+++ b/clip/clip.py
@@ -17,6 +17,8 @@ _tokenizer = _Tokenizer()
 
 _MODELS = {
     "RN50": "https://openaipublic.azureedge.net/clip/models/afeb0e10f9e5a86da6080e35cf09123aca3b358a0c3e3b6c78a7b63bc04b6762/RN50.pt",
+    "RN101": "https://openaipublic.azureedge.net/clip/models/8fa8567bab74a42d41c5915025a8e4538c3bdbe8804a470a72f30b0d94fab599/RN101.pt",
+    "RN50x4": "https://openaipublic.azureedge.net/clip/models/7e526bd135e493cef0776de27d5f42653e6b4c8bf9e0f653bb11773263205fdd/RN50x4.pt",
     "ViT-B/32": "https://openaipublic.azureedge.net/clip/models/40d365715913c9da98579312b702a82c18be219cc2a73407c4526f58eba950af/ViT-B-32.pt",
 }
 
diff --git a/model-card.md b/model-card.md
index abb3031..003e7e7 100644
--- a/model-card.md
+++ b/model-card.md
@@ -16,14 +16,16 @@ The base model uses a ResNet50 with several modifications as an image encoder an
 
 ### Model Version
 
-Initially we’ve released one CLIP model based on the Vision Transformer architecture equivalent to ViT-B/32
+Initially, we’ve released one CLIP model based on the Vision Transformer architecture equivalent to ViT-B/32, along with the RN50 model, using the architecture equivalent to ResNet-50.
+
+As part of the staged release process, we have also released the RN101 model, as well as RN50x4, a RN50 scaled up 4x according to the [EfficientNet](https://arxiv.org/abs/1905.11946) scaling rule.
 
 Please see the paper linked below for further details about their specification.
 
 ### Documents
 
 - [Blog Post](https://openai.com/blog/clip/)
-- [CLIP Paper](https://cdn.openai.com/papers/Learning_Transferable_Visual_Models_From_Natural_Language_Supervision.pdf)
+- [CLIP Paper](https://arxiv.org/abs/2103.00020)