crfm-helm 0.4.0__py3-none-any.whl → 0.5.1__py3-none-any.whl

helm/config/model_metadata.yaml

@@ -19,6 +19,17 @@ models:
     release_date: 2023-01-01
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
 
+  # Adobe
+  - name: adobe/giga-gan
+    display_name: GigaGAN (1B)
+    description: GigaGAN is a GAN model that produces high-quality images extremely quickly. The model was trained on text and image pairs from LAION2B-en and COYO-700M. ([paper](https://arxiv.org/abs/2303.05511)).
+    creator_organization_name: Adobe
+    access: limited
+    num_parameters: 1000000000
+    release_date: 2023-06-22
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+
   # AI21 Labs
   - name: ai21/j1-jumbo # DEPRECATED
     display_name: J1-Jumbo v1 (178B)
@@ -102,7 +113,7 @@ models:
     # TODO: get exact release date
     release_date: 2022-01-01
     # Does not support echo
-    tags: [TEXT_MODEL_TAG, IMAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   - name: AlephAlpha/luminous-extended
     display_name: Luminous Extended (30B)
@@ -112,7 +123,7 @@ models:
     num_parameters: 30000000000
     release_date: 2022-01-01
     # Does not support echo
-    tags: [TEXT_MODEL_TAG, IMAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   - name: AlephAlpha/luminous-supreme
     display_name: Luminous Supreme (70B)
@@ -122,7 +133,7 @@ models:
     num_parameters: 70000000000
     release_date: 2022-01-01
     # Does not support echo.
-    # TODO: images will be supported in the near future. Add IMAGE_MODEL_TAG.
+    # Currently, only Luminous-extended and Luminous-base support multimodal inputs
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
   
   # TODO: Uncomment when luminous-world is released.
@@ -135,12 +146,50 @@ models:
   #   release_date: TBD
   #   # Does not support echo.
   #   tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+  
+  - name: AlephAlpha/m-vader
+    display_name: MultiFusion (13B)
+    description: MultiFusion is a multimodal, multilingual diffusion model that extend the capabilities of Stable Diffusion v1.4 by integrating different pre-trained modules, which transfers capabilities to the downstream model ([paper](https://arxiv.org/abs/2305.15296))
+    creator_organization_name: Aleph Alpha
+    access: limited
+    num_parameters: 13000000000
+    release_date: 2023-05-24
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
 
+  # Amazon
+  # References for Amazon Titan models:
+  # - https://aws.amazon.com/bedrock/titan/
+  # - https://community.aws/content/2ZUVD3fkNtqEOYIa2iUJAFArS7c/family-of-titan-text-models---cli-demo
+  # - https://aws.amazon.com/about-aws/whats-new/2023/11/amazon-titan-models-express-lite-bedrock/
+  - name: amazon/titan-text-lite-v1
+    display_name: Amazon Titan Text Lite
+    description: Amazon Titan Text Lite is a lightweight, efficient model perfect for fine-tuning English-language tasks like summarization and copywriting. It caters to customers seeking a smaller, cost-effective, and highly customizable model. It supports various formats, including text generation, code generation, rich text formatting, and orchestration (agents). Key model attributes encompass fine-tuning, text generation, code generation, and rich text formatting.
+    creator_organization_name: Amazon
+    access: limited
+    release_date: 2023-11-29
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: amazon/titan-tg1-large
+    display_name: Amazon Titan Large
+    description: Amazon Titan Large is efficient model perfect for fine-tuning English-language tasks like summarization, create article, marketing campaign.
+    creator_organization_name: Amazon
+    access: limited
+    release_date: 2023-11-29
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: amazon/titan-text-express-v1
+    display_name: Amazon Titan Text Express
+    description: Amazon Titan Text Express, with a context length of up to 8,000 tokens, excels in advanced language tasks like open-ended text generation and conversational chat. It's also optimized for Retrieval Augmented Generation (RAG). Initially designed for English, the model offers preview multilingual support for over 100 additional languages.
+    creator_organization_name: Amazon
+    access: limited
+    release_date: 2023-11-29
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
 
 
   # Anthropic
   - name: anthropic/claude-v1.3
-    display_name: Anthropic Claude v1.3
+    display_name: Claude v1.3
     description: A 52B parameter language model, trained using reinforcement learning from human feedback [paper](https://arxiv.org/pdf/2204.05862.pdf).
     creator_organization_name: Anthropic
     access: limited
@@ -149,7 +198,7 @@ models:
     tags: [ANTHROPIC_CLAUDE_1_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
  
   - name: anthropic/claude-instant-v1
-    display_name: Anthropic Claude Instant V1
+    display_name: Claude Instant V1
     description: A lightweight version of Claude, a model trained using reinforcement learning from human feedback ([docs](https://www.anthropic.com/index/introducing-claude)).
     creator_organization_name: Anthropic
     access: limited
@@ -157,7 +206,7 @@ models:
     tags: [ANTHROPIC_CLAUDE_1_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: anthropic/claude-instant-1.2
-    display_name: Anthropic Claude Instant 1.2
+    display_name: Claude Instant 1.2
     description: A lightweight version of Claude, a model trained using reinforcement learning from human feedback ([docs](https://www.anthropic.com/index/introducing-claude)).
     creator_organization_name: Anthropic
     access: limited
@@ -165,7 +214,7 @@ models:
     tags: [ANTHROPIC_CLAUDE_1_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: anthropic/claude-2.0
-    display_name: Anthropic Claude 2.0
+    display_name: Claude 2.0
     description: Claude 2.0 is a general purpose large language model developed by Anthropic. It uses a transformer architecture and is trained via unsupervised learning, RLHF, and Constitutional AI (including both a supervised and Reinforcement Learning (RL) phase). ([model card](https://efficient-manatee.files.svdcdn.com/production/images/Model-Card-Claude-2.pdf))
     creator_organization_name: Anthropic
     access: limited
@@ -173,13 +222,37 @@ models:
     tags: [ANTHROPIC_CLAUDE_2_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: anthropic/claude-2.1
-    display_name: Anthropic Claude 2.1
+    display_name: Claude 2.1
     description: Claude 2.1 is a general purpose large language model developed by Anthropic. It uses a transformer architecture and is trained via unsupervised learning, RLHF, and Constitutional AI (including both a supervised and Reinforcement Learning (RL) phase). ([model card](https://efficient-manatee.files.svdcdn.com/production/images/Model-Card-Claude-2.pdf))
     creator_organization_name: Anthropic
     access: limited
     release_date: 2023-11-21
     tags: [ANTHROPIC_CLAUDE_2_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: anthropic/claude-3-haiku-20240307
+    display_name: Claude 3 Haiku (20240307)
+    description: Claude 3 is a a family of models that possess vision and multilingual capabilities. They were trained with various methods such as unsupervised learning and Constitutional AI ([blog](https://www.anthropic.com/news/claude-3-family)).
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2024-03-13  # https://www.anthropic.com/news/claude-3-haiku
+    tags: [ANTHROPIC_CLAUDE_3_MODEL_TAG, TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-3-sonnet-20240229
+    display_name: Claude 3 Sonnet (20240229)
+    description: Claude 3 is a a family of models that possess vision and multilingual capabilities. They were trained with various methods such as unsupervised learning and Constitutional AI ([blog](https://www.anthropic.com/news/claude-3-family)).
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2024-03-04  # https://www.anthropic.com/news/claude-3-family
+    tags: [ANTHROPIC_CLAUDE_3_MODEL_TAG, TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-3-opus-20240229
+    display_name: Claude 3 Opus (20240229)
+    description: Claude 3 is a a family of models that possess vision and multilingual capabilities. They were trained with various methods such as unsupervised learning and Constitutional AI ([blog](https://www.anthropic.com/news/claude-3-family)).
+    access: limited
+    creator_organization_name: Anthropic
+    release_date: 2024-03-04  # https://www.anthropic.com/news/claude-3-family
+    tags: [ANTHROPIC_CLAUDE_3_MODEL_TAG, TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   # DEPRECATED: Please do not use.
   - name: anthropic/stanford-online-all-v4-s3
     display_name: Anthropic-LM v4-s3 (52B)
@@ -378,6 +451,52 @@ models:
     release_date: 2023-09-29
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  # Craiyon
+  - name: craiyon/dalle-mini
+    display_name: DALL-E mini (0.4B)
+    description: DALL-E mini is an open-source text-to-image model that attempt to reproduce OpenAI's DALL-E 1 ([code](https://github.com/borisdayma/dalle-mini)).
+    creator_organization_name: Craiyon
+    access: open
+    num_parameters: 400000000
+    release_date: 2022-04-21
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: craiyon/dalle-mega
+    display_name: DALL-E mega (2.6B)
+    description: DALL-E mega is an open-source text-to-image model that attempt to reproduce OpenAI's DALL-E 1 ([code](https://github.com/borisdayma/dalle-mini)).
+    creator_organization_name: Craiyon
+    access: open
+    num_parameters: 2600000000
+    release_date: 2022-04-21
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  # DeepFloyd
+  - name: DeepFloyd/IF-I-M-v1.0
+    display_name: DeepFloyd IF Medium (0.4B)
+    description: DeepFloyd-IF is a pixel-based text-to-image triple-cascaded diffusion model with state-of-the-art photorealism and language understanding (paper coming soon).
+    creator_organization_name: DeepFloyd
+    access: open
+    num_parameters: 400000000
+    release_date: 2023-04-28
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: DeepFloyd/IF-I-L-v1.0
+    display_name: DeepFloyd IF Large (0.9B)
+    description: DeepFloyd-IF is a pixel-based text-to-image triple-cascaded diffusion model with state-of-the-art photorealism and language understanding (paper coming soon).
+    creator_organization_name: DeepFloyd
+    access: open
+    num_parameters: 900000000
+    release_date: 2023-04-28
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: DeepFloyd/IF-I-XL-v1.0
+    display_name: DeepFloyd IF X-Large (4.3B)
+    description: DeepFloyd-IF is a pixel-based text-to-image triple-cascaded diffusion model with state-of-the-art photorealism and language understanding (paper coming soon).
+    creator_organization_name: DeepFloyd
+    access: open
+    num_parameters: 4300000000
+    release_date: 2023-04-28
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
 
 
   # Databricks
@@ -408,6 +527,14 @@ models:
     release_date: 2023-04-12
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
 
+  - name: databricks/dbrx-instruct
+    display_name: DBRX Instruct
+    description: DBRX is a large language model with a fine-grained mixture-of-experts (MoE) architecture that uses 16 experts and chooses 4. It has 132B total parameters, of which 36B parameters are active on any input. ([blog post](https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm))
+    creator_organization_name: Databricks
+    access: open
+    num_parameters: 132000000000
+    release_date: 2024-03-27
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
 
   # DeepMind
@@ -430,6 +557,16 @@ models:
     tags: [] # TODO: add tags
 
 
+  # Deepseek
+  - name: deepseek-ai/deepseek-llm-67b-chat
+    display_name: DeepSeek LLM Chat (67B)
+    description: DeepSeek LLM Chat is a open-source language model trained on 2 trillion tokens in both English and Chinese, and fine-tuned supervised fine-tuning (SFT) and Direct Preference Optimization (DPO). ([paper](https://arxiv.org/abs/2401.02954))
+    creator_organization_name: DeepSeek
+    access: open
+    num_parameters: 67000000000
+    release_date: 2024-01-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
 
   # EleutherAI
   - name: eleutherai/gpt-j-6b # Served by GooseAi, HuggingFace and Together.
@@ -526,6 +663,85 @@ models:
     release_date: 2023-03-01 # was first announced on 2022-04 but remained private.
     tags: [] # TODO: add tags
 
+    # Note: This is aliased to a snapshot of gemini-pro. When possible, please use a versioned snapshot instead.
+  - name: google/gemini-pro
+    display_name: Gemini Pro
+    description: Gemini Pro is a multimodal model able to reason across text, images, video, audio and code. ([paper](https://arxiv.org/abs/2312.11805))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-12-13
+    tags: [TEXT_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-1.0-pro-001
+    display_name: Gemini 1.0 Pro
+    description: Gemini 1.0 Pro is a multimodal model able to reason across text, images, video, audio and code. ([paper](https://arxiv.org/abs/2312.11805))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-12-13
+    tags: [TEXT_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+    # Note: This is aliased to a snapshot of gemini-pro-vision. When possible, please use a versioned snapshot instead.
+  - name: google/gemini-pro-vision
+    display_name: Gemini Pro Vision
+    description: Gemini Pro Vision is a multimodal model able to reason across text, images, video, audio and code. ([paper](https://arxiv.org/abs/2312.11805))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-12-13
+    tags: [VISION_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG]
+
+  - name: google/gemini-1.0-pro-vision-001
+    display_name: Gemini 1.0 Pro Vision
+    description: Gemini 1.0 Pro Vision is a multimodal model able to reason across text, images, video, audio and code. ([paper](https://arxiv.org/abs/2312.11805))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-12-13
+    tags: [VISION_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, GOOGLE_GEMINI_PRO_VISION_V1_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: google/gemini-1.5-pro-preview-0409
+    display_name: Gemini 1.5 Pro (0409 preview)
+    description: Gemini 1.5 Pro is a multimodal mixture-of-experts model capable of recalling and reasoning over fine-grained information from long contexts. ([paper](https://arxiv.org/abs/2403.05530))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2024-04-10
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemma-2b
+    display_name: Gemma (2B)
+    # TODO: Fill in Gemma description.
+    description: TBD
+    creator_organization_name: Google
+    access: open
+    release_date: 2024-02-21
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: google/gemma-2b-it
+    display_name: Gemma Instruct (2B)
+    # TODO: Fill in Gemma description.
+    description: TBD
+    creator_organization_name: Google
+    access: open
+    release_date: 2024-02-21
+    tags: [TEXT_MODEL_TAG, GOOGLE_GEMMA_INSTRUCT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemma-7b
+    display_name: Gemma (7B)
+    # TODO: Fill in Gemma description.
+    description: TBD
+    creator_organization_name: Google
+    access: open
+    release_date: 2024-02-21
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: google/gemma-7b-it
+    display_name: Gemma Instruct (7B)
+    # TODO: Fill in Gemma description.
+    description: TBD
+    creator_organization_name: Google
+    access: open
+    release_date: 2024-02-21
+    # TODO: Add OUTPUT_FORMAT_INSTRUCTIONS_TAG tag
+    tags: [TEXT_MODEL_TAG, GOOGLE_GEMMA_INSTRUCT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: google/text-bison@001
     display_name: PaLM-2 (Bison)
     description: The best value PaLM model. PaLM 2 (Pathways Language Model) is a Transformer-based model trained using a mixture of objectives that was evaluated on English and multilingual language, and reasoning tasks. ([report](https://arxiv.org/pdf/2305.10403.pdf))
@@ -534,6 +750,14 @@ models:
     release_date: 2023-06-07 # Source: https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/text#model_versions
     tags: [TEXT_MODEL_TAG, GOOGLE_PALM_2_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
 
+  - name: google/text-bison@002
+    display_name: PaLM-2 (Bison)
+    description: The best value PaLM model. PaLM 2 (Pathways Language Model) is a Transformer-based model trained using a mixture of objectives that was evaluated on English and multilingual language, and reasoning tasks. ([report](https://arxiv.org/pdf/2305.10403.pdf))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-06-07 # Source: https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/text#model_versions
+    tags: [TEXT_MODEL_TAG, GOOGLE_PALM_2_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
   - name: google/text-bison-32k
     display_name: PaLM-2 (Bison)
     description: The best value PaLM model with a 32K context. PaLM 2 (Pathways Language Model) is a Transformer-based model trained using a mixture of objectives that was evaluated on English and multilingual language, and reasoning tasks. ([report](https://arxiv.org/pdf/2305.10403.pdf))
@@ -558,6 +782,14 @@ models:
     release_date: 2023-06-29 # Source: https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/code-generation#model_versions
     tags: [CODE_MODEL_TAG]
 
+  - name: google/code-bison@002
+    display_name: Codey PaLM-2 (Bison)
+    description: A model fine-tuned to generate code based on a natural language description of the desired code. PaLM 2 (Pathways Language Model) is a Transformer-based model trained using a mixture of objectives that was evaluated on English and multilingual language, and reasoning tasks. ([report](https://arxiv.org/pdf/2305.10403.pdf))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2023-06-29 # Source: https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/code-generation#model_versions
+    tags: [CODE_MODEL_TAG]
+
   - name: google/code-bison-32k
     display_name: Codey PaLM-2 (Bison)
     description: Codey with a 32K context. PaLM 2 (Pathways Language Model) is a Transformer-based model trained using a mixture of objectives that was evaluated on English and multilingual language, and reasoning tasks. ([report](https://arxiv.org/pdf/2305.10403.pdf))
@@ -568,55 +800,233 @@ models:
 
 
 
-  # HazyResearch
-  - name: hazyresearch/h3-2.7b
-    display_name: H3 (2.7B)
-    description: H3 (2.7B parameters) is a decoder-only language model based on state space models ([paper](https://arxiv.org/abs/2212.14052)).
-    creator_organization_name: HazyResearch
+  # HuggingFace
+  - name: HuggingFaceM4/idefics2-8b
+    display_name: IDEFICS 2 (8B)
+    description: IDEFICS 2 (8B parameters) is an open multimodal model that accepts arbitrary sequences of image and text inputs and produces text outputs. ([blog](https://huggingface.co/blog/idefics2)).
+    creator_organization_name: HuggingFace
     access: open
-    num_parameters: 2700000000
-    release_date: 2023-01-23
-    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+    num_parameters: 8000000000
+    release_date: 2024-04-15
+    tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
-
-
-  # HuggingFace
   - name: HuggingFaceM4/idefics-9b
     display_name: IDEFICS (9B)
-    description: IDEFICS (9B parameters) is an open-source model based on DeepMind's Flamingo. ([blog](https://huggingface.co/blog/idefics))
+    description: IDEFICS (9B parameters) is an open-source model based on DeepMind's Flamingo ([blog](https://huggingface.co/blog/idefics)).
     creator_organization_name: HuggingFace
     access: open
     num_parameters: 9000000000
     release_date: 2023-08-22
-    tags: [VISION_LANGUAGE_MODEL_TAG]
+    tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   - name: HuggingFaceM4/idefics-9b-instruct
-    display_name: IDEFICS instruct (9B)
-    description: IDEFICS instruct (9B parameters) is an open-source model based on DeepMind's Flamingo. ([blog](https://huggingface.co/blog/idefics))
+    display_name: IDEFICS-instruct (9B)
+    description: IDEFICS-instruct (9B parameters) is the instruction-tuned version of IDEFICS 9B ([blog](https://huggingface.co/blog/idefics)).
     creator_organization_name: HuggingFace
     access: open
     num_parameters: 9000000000
     release_date: 2023-08-22
-    tags: [VISION_LANGUAGE_MODEL_TAG]
+    tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, IDEFICS_INSTRUCT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   - name: HuggingFaceM4/idefics-80b
     display_name: IDEFICS (80B)
-    description: IDEFICS (80B parameters) is an open-source model based on DeepMind's Flamingo. ([blog](https://huggingface.co/blog/idefics))
+    description: IDEFICS (80B parameters) is an open-source model based on DeepMind's Flamingo ([blog](https://huggingface.co/blog/idefics)).
     creator_organization_name: HuggingFace
     access: open
     num_parameters: 80000000000
     release_date: 2023-08-22
-    tags: [VISION_LANGUAGE_MODEL_TAG]
+    tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   - name: HuggingFaceM4/idefics-80b-instruct
-    display_name: IDEFICS instruct (80B)
-    description: IDEFICS instruct (80B parameters) is an open-source model based on DeepMind's Flamingo. ([blog](https://huggingface.co/blog/idefics))
+    display_name: IDEFICS-instruct (80B)
+    description: IDEFICS-instruct (80B parameters) is the instruction-tuned version of IDEFICS 80B ([blog](https://huggingface.co/blog/idefics)).
     creator_organization_name: HuggingFace
     access: open
     num_parameters: 80000000000
     release_date: 2023-08-22
-    tags: [VISION_LANGUAGE_MODEL_TAG]
+    tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, IDEFICS_INSTRUCT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  ## Text-to-Image Diffusion Models
+  - name: huggingface/dreamlike-diffusion-v1-0
+    display_name: Dreamlike Diffusion v1.0 (1B)
+    description: Dreamlike Diffusion v1.0 is Stable Diffusion v1.5 fine tuned on high quality art ([HuggingFace model card](https://huggingface.co/dreamlike-art/dreamlike-diffusion-1.0))
+    creator_organization_name: dreamlike.art
+    access: open
+    num_parameters: 1000000000
+    release_date: 2023-03-08
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/dreamlike-photoreal-v2-0
+    display_name: Dreamlike Photoreal v2.0 (1B)
+    description: Dreamlike Photoreal v2.0 is a photorealistic model based on Stable Diffusion v1.5 ([HuggingFace model card](https://huggingface.co/dreamlike-art/dreamlike-photoreal-2.0))
+    creator_organization_name: dreamlike.art
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-23
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/openjourney-v1-0
+    display_name: Openjourney (1B)
+    description: Openjourney is an open source Stable Diffusion fine tuned model on Midjourney images ([HuggingFace model card](https://huggingface.co/prompthero/openjourney))
+    creator_organization_name: PromptHero
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-01  # TODO: get the exact date
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/openjourney-v2-0
+    display_name: Openjourney v2 (1B)
+    description: Openjourney v2 is an open source Stable Diffusion fine tuned model on Midjourney images. Openjourney v2 is now referred to as Openjourney v4 in Hugging Face ([HuggingFace model card](https://huggingface.co/prompthero/openjourney-v4)).
+    creator_organization_name: PromptHero
+    access: open
+    num_parameters: 1000000000
+    release_date: 2023-01-01  # TODO: get the exact date
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/promptist-stable-diffusion-v1-4
+    display_name: Promptist + Stable Diffusion v1.4 (1B)
+    description: Trained with human preferences, Promptist optimizes user input into model-preferred prompts for Stable Diffusion v1.4 ([paper](https://arxiv.org/abs/2212.09611))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-12-19
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/redshift-diffusion
+    display_name: Redshift Diffusion (1B)
+    description: Redshift Diffusion is an open source Stable Diffusion model fine tuned on high resolution 3D artworks ([HuggingFace model card](https://huggingface.co/nitrosocke/redshift-diffusion))
+    creator_organization_name: nitrosocke
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-29
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-safe-weak
+    display_name: Safe Stable Diffusion weak (1B)
+    description: Safe Stable Diffusion is an extension to the Stable Diffusion that drastically reduces inappropriate content ([paper](https://arxiv.org/abs/2211.05105)).
+    creator_organization_name: TU Darmstadt
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-09
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-safe-medium
+    display_name: Safe Stable Diffusion medium (1B)
+    description: Safe Stable Diffusion is an extension to the Stable Diffusion that drastically reduces inappropriate content ([paper](https://arxiv.org/abs/2211.05105))
+    creator_organization_name: TU Darmstadt
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-09
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-safe-strong
+    display_name: Safe Stable Diffusion strong (1B)
+    description: Safe Stable Diffusion is an extension to the Stable Diffusion that drastically reduces inappropriate content ([paper](https://arxiv.org/abs/2211.05105))
+    creator_organization_name: TU Darmstadt
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-09
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-safe-max
+    display_name: Safe Stable Diffusion max (1B)
+    description: Safe Stable Diffusion is an extension to the Stable Diffusion that drastically reduces inappropriate content ([paper](https://arxiv.org/abs/2211.05105))
+    creator_organization_name: TU Darmstadt
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-09
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-v1-4
+    display_name: Stable Diffusion v1.4 (1B)
+    description: Stable Diffusion v1.4 is a latent text-to-image diffusion model capable of generating photorealistic images given any text input ([paper](https://arxiv.org/abs/2112.10752))
+    creator_organization_name: Ludwig Maximilian University of Munich CompVis
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-08-01
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-v1-5
+    display_name: Stable Diffusion v1.5 (1B)
+    description: The Stable-Diffusion-v1-5 checkpoint was initialized with the weights of the Stable-Diffusion-v1-2 checkpoint and subsequently fine-tuned on 595k steps at resolution 512x512 on laion-aesthetics v2 5+ and 10% dropping of the text-conditioning to improve classifier-free guidance sampling ([paper](https://arxiv.org/abs/2112.10752))
+    creator_organization_name: Runway
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-10-20
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-v2-base
+    display_name: Stable Diffusion v2 base (1B)
+    description: The model is trained from scratch 550k steps at resolution 256x256 on a subset of LAION-5B filtered for explicit pornographic material, using the LAION-NSFW classifier with punsafe=0.1 and an aesthetic score greater than 4.5. Then it is further trained for 850k steps at resolution 512x512 on the same dataset on images with resolution greater than 512x512 ([paper](https://arxiv.org/abs/2112.10752))
+    creator_organization_name: Stability AI
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-23
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/stable-diffusion-v2-1-base
+    display_name: Stable Diffusion v2.1 base (1B)
+    description: This stable-diffusion-2-1-base model fine-tunes stable-diffusion-2-base with 220k extra steps taken, with punsafe=0.98 on the same dataset ([paper](https://arxiv.org/abs/2112.10752))
+    creator_organization_name: Stability AI
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-11-23
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: huggingface/vintedois-diffusion-v0-1
+    display_name: Vintedois (22h) Diffusion model v0.1 (1B)
+    description: Vintedois (22h) Diffusion model v0.1 is Stable Diffusion v1.5 that was finetuned on a large amount of high quality images with simple prompts to generate beautiful images without a lot of prompt engineering ([HuggingFace model card](https://huggingface.co/22h/vintedois-diffusion-v0-1))
+    creator_organization_name: 22 Hours
+    access: open
+    num_parameters: 1000000000
+    release_date: 2022-12-27
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
 
+  - name: segmind/Segmind-Vega
+    display_name: Segmind Stable Diffusion (0.74B)
+    description: The Segmind-Vega Model is a distilled version of the Stable Diffusion XL (SDXL), offering a remarkable 70% reduction in size and an impressive 100% speedup while retaining high-quality text-to-image generation capabilities. Trained on diverse datasets, including Grit and Midjourney scrape data, it excels at creating a wide range of visual content based on textual prompts. ([HuggingFace model card](https://huggingface.co/segmind/Segmind-Vega))
+    creator_organization_name: Segmind
+    access: open
+    num_parameters: 740000000
+    release_date: 2023-12-01
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: segmind/SSD-1B
+    display_name: Segmind Stable Diffusion (1B)
+    description: The Segmind Stable Diffusion Model (SSD-1B) is a distilled 50% smaller version of the Stable Diffusion XL (SDXL), offering a 60% speedup while maintaining high-quality text-to-image generation capabilities. It has been trained on diverse datasets, including Grit and Midjourney scrape data, to enhance its ability to create a wide range of visual content based on textual prompts. ([HuggingFace model card](https://huggingface.co/segmind/SSD-1B))
+    creator_organization_name: Segmind
+    access: open
+    num_parameters: 1000000000
+    release_date: 2023-10-20
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: stabilityai/stable-diffusion-xl-base-1.0
+    display_name: Stable Diffusion XL
+    description: Stable Diffusion XL (SDXL) consists of an ensemble of experts pipeline for latent diffusion. ([HuggingFace model card](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0))
+    creator_organization_name: Stability AI
+    access: open
+    num_parameters: 6600000000
+    release_date: 2023-07-26
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  # Kakao
+  - name: kakaobrain/mindall-e
+    display_name: minDALL-E (1.3B)
+    description: minDALL-E, named after minGPT, is an autoregressive text-to-image generation model trained on 14 million image-text pairs ([code](https://github.com/kakaobrain/minDALL-E))
+    creator_organization_name: Kakao
+    access: open
+    num_parameters: 1300000000
+    release_date: 2021-12-13
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  # Lexica
+  - name: lexica/search-stable-diffusion-1.5
+    display_name: Lexica Search with Stable Diffusion v1.5 (1B)
+    description: Retrieves Stable Diffusion v1.5 images Lexica users generated ([docs](https://lexica.art/docs)).
+    creator_organization_name: Lexica
+    access: open
+    release_date: 2023-01-01
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
 
 
   # Lightning AI
@@ -791,6 +1201,41 @@ models:
     # TODO(#1828): Upgrade to FULL_FUNCTIONALITY_TEXT_MODEL_TAG
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
 
+  - name: meta/llama-3-8b
+    display_name: Llama 3 (8B)
+    description: Llama 3 is a family of language models that have been trained on more than 15 trillion tokens, and use Grouped-Query Attention (GQA) for improved inference scalability.
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 8000000000
+    release_date: 2024-04-18
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+    
+  - name: meta/llama-3-70b
+    display_name: Llama 3 (70B)
+    description: Llama 3 is a family of language models that have been trained on more than 15 trillion tokens, and use Grouped-Query Attention (GQA) for improved inference scalability.
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 70000000000
+    release_date: 2024-04-18
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: meta/llama-3-8b-chat
+    display_name: Llama 3 Chat (8B)
+    description: Llama 3 is a family of language models that have been trained on more than 15 trillion tokens, and use Grouped-Query Attention (GQA) for improved inference scalability. It used SFT, rejection sampling, PPO and DPO for post-training.
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 8000000000
+    release_date: 2024-04-18
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: meta/llama-3-70b-chat
+    display_name: Llama 3 Chat (70B)
+    description: Llama 3 is a family of language models that have been trained on more than 15 trillion tokens, and use Grouped-Query Attention (GQA) for improved inference scalability. It used SFT, rejection sampling, PPO and DPO for post-training.
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 70000000000
+    release_date: 2024-04-18
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
 
   # Microsoft/NVIDIA
@@ -801,7 +1246,7 @@ models:
     access: closed
     num_parameters: 530000000000
     release_date: 2022-01-28
-    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+    tags: []  # deprecated text model
 
   - name: microsoft/TNLGv2_7B
     display_name: TNLG v2 (6.7B)
@@ -810,7 +1255,79 @@ models:
     access: closed
     num_parameters: 6700000000
     release_date: 2022-01-28
-    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+    tags: []  # deprecated text model
+
+  - name: microsoft/llava-1.5-7b-hf
+    display_name: LLaVA 1.5 (7B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 7000000000
+    release_date: 2023-10-05
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: microsoft/llava-1.5-13b-hf
+    display_name: LLaVA 1.5 (13B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 13000000000
+    release_date: 2023-10-05
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: uw-madison/llava-v1.6-vicuna-7b-hf
+    display_name: LLaVA 1.6 (7B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-01-01
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: uw-madison/llava-v1.6-vicuna-13b-hf
+    display_name: LLaVA 1.6 (13B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 13000000000
+    release_date: 2024-01-01
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: uw-madison/llava-v1.6-mistral-7b-hf
+    display_name: LLaVA 1.6 + Mistral (7B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-01-01
+    tags: [ VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG ]
+
+  - name: uw-madison/llava-v1.6-34b-hf
+    display_name: LLaVA + Nous-Hermes-2-Yi-34B (34B)
+    description: LLaVa is an open-source chatbot trained by fine-tuning LlamA/Vicuna on GPT-generated multimodal instruction-following data. ([paper](https://arxiv.org/abs/2304.08485))
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 34000000000
+    release_date: 2024-01-01
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+  
+  - name: openflamingo/OpenFlamingo-9B-vitl-mpt7b
+    display_name: OpenFlamingo (9B)
+    description: OpenFlamingo is an open source implementation of DeepMind's Flamingo models. This 9B-parameter model uses a CLIP ViT-L/14 vision encoder and MPT-7B language model ([paper](https://arxiv.org/abs/2308.01390)).
+    creator_organization_name: OpenFlamingo
+    access: open
+    num_parameters: 9000000000
+    release_date: 2023-08-02
+    tags: [VISION_LANGUAGE_MODEL_TAG, OPEN_FLAMINGO_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: microsoft/phi-2
+    display_name: Phi-2
+    description: Phi-2 is a Transformer with 2.7 billion parameters. It was trained using the same data sources as Phi-1.5, augmented with a new data source that consists of various NLP synthetic texts and filtered websites (for safety and educational value)
+    creator_organization_name: Microsoft
+    access: open
+    num_parameters: 13000000000
+    release_date: 2023-10-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
 
 
 
@@ -831,6 +1348,52 @@ models:
     num_parameters: 34000000000
     release_date: 2023-11-02
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+  - name: 01-ai/yi-6b-chat
+    display_name: Yi Chat (6B)
+    description: The Yi models are large language models trained from scratch by developers at 01.AI.
+    creator_organization_name: 01.AI
+    access: open
+    num_parameters: 6000000000
+    release_date: 2023-11-23
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+  - name: 01-ai/yi-34b-chat
+    display_name: Yi Chat (34B)
+    description: The Yi models are large language models trained from scratch by developers at 01.AI.
+    creator_organization_name: 01.AI
+    access: open
+    num_parameters: 34000000000
+    release_date: 2023-11-23
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  # Allen Institute for AI
+  # OLMo Blog: https://blog.allenai.org/olmo-open-language-model-87ccfc95f580
+  - name: allenai/olmo-7b
+    display_name: OLMo (7B)
+    description: OLMo is a series of Open Language Models trained on the Dolma dataset.
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-02-01
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: allenai/olmo-7b-twin-2t
+    display_name: OLMo (7B Twin 2T)
+    description: OLMo is a series of Open Language Models trained on the Dolma dataset.
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-02-01
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: allenai/olmo-7b-instruct
+    display_name: OLMo (7B Instruct)
+    description: OLMo is a series of Open Language Models trained on the Dolma dataset. The instruct versions was trained on the Tulu SFT mixture and a cleaned version of the UltraFeedback dataset.
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-02-01
+    # TODO: Add instruct tag.
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
 
 
   # Mistral AI
@@ -841,17 +1404,81 @@ models:
     access: open
     num_parameters: 7300000000
     release_date: 2023-09-27
-    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: mistralai/mixtral-8x7b-32kseqlen
     display_name: Mixtral (8x7B 32K seqlen)
     description: Mistral AI's mixture-of-experts model ([tweet](https://twitter.com/MistralAI/status/1733150512395038967)).
     creator_organization_name: Mistral AI
     access: open
-    num_parameters: 56000000000
+    num_parameters: 46700000000
     release_date: 2023-12-08
-    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: mistralai/mixtral-8x7b-instruct-v0.1
+    display_name: Mixtral (8x7B Instruct)
+    description: Mixtral (8x7B Instruct) is a version of Mixtral (8x7B) that was optimized through supervised fine-tuning and direct preference optimisation (DPO) for careful instruction following.
+    creator_organization_name: Mistral AI
+    access: open
+    num_parameters: 46700000000
+    # Blog post: https://mistral.ai/news/mixtral-of-experts/
+    release_date: 2023-12-11
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG, MISTRAL_MODEL_TAG]
+
+  - name: mistralai/mixtral-8x22b
+    display_name: Mixtral (8x22B)
+    description: Mistral AI's mixture-of-experts model that uses 39B active parameters out of 141B ([blog post](https://mistral.ai/news/mixtral-8x22b/)).
+    creator_organization_name: Mistral AI
+    access: open
+    num_parameters: 176000000000
+    release_date: 2024-04-10
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: mistralai/mixtral-8x22b-instruct-v0.1
+    display_name: Mixtral Instruct (8x22B)
+    description: Mistral AI's mixture-of-experts model that uses 39B active parameters out of 141B ([blog post](https://mistral.ai/news/mixtral-8x22b/)).
+    creator_organization_name: Mistral AI
+    access: open
+    num_parameters: 176000000000
+    release_date: 2024-04-10
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: mistralai/bakLlava-v1-hf
+    display_name: BakLLaVA v1 (7B)
+    description: BakLLaVA v1 is a Mistral 7B base augmented with the LLaVA 1.5 architecture. ([blog](https://huggingface.co/llava-hf/bakLlava-v1-hf))
+    creator_organization_name: Mistral AI
+    access: open
+    num_parameters: 7000000000
+    release_date: 2023-10-16
+    tags: [VISION_LANGUAGE_MODEL_TAG, LLAVA_MODEL_TAG, LIMITED_FUNCTIONALITY_VLM_TAG]
+
+  - name: mistralai/mistral-small-2402
+    display_name: Mistral Small (2402)
+    # TODO: Fill in description
+    description: TBD
+    creator_organization_name: Mistral AI
+    access: limited
+    # Blog post: https://mistral.ai/news/mistral-large/
+    release_date: 2023-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG, MISTRAL_MODEL_TAG]
+
+  - name: mistralai/mistral-medium-2312
+    display_name: Mistral Medium (2312)
+    description: Mistral is a transformer model that uses Grouped-Query Attention (GQA) and Sliding-Window Attention (SWA).
+    creator_organization_name: Mistral AI
+    access: limited
+    release_date: 2023-12-11
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG, MISTRAL_MODEL_TAG]
+
+  - name: mistralai/mistral-large-2402
+    display_name: Mistral Large (2402)
+    # TODO: Fill in description
+    description: TBD
+    creator_organization_name: Mistral AI
+    access: limited
+    # Blog post: https://mistral.ai/news/mistral-large/
+    release_date: 2023-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG, MISTRAL_MODEL_TAG]
 
 
   # MosaicML
@@ -950,6 +1577,23 @@ models:
 
   ## GPT 3 Models
   # The list of models can be found here: https://beta.openai.com/docs/engines/gpt-3
+
+  - name: openai/davinci-002
+    display_name: davinci-002
+    description: Replacement for the GPT-3 curie and davinci base models.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-08-22
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: openai/babbage-002
+    display_name: babbage-002
+    description: Replacement for the GPT-3 ada and babbage base models.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-08-22
+    tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
   # DEPRECATED: Announced on July 06 2023 that these models will be shut down on January 04 2024.
 
   - name: openai/davinci # DEPRECATED
@@ -1048,9 +1692,17 @@ models:
   ## GPT 3.5 Turbo Models
   # ChatGPT: https://openai.com/blog/chatgpt
   
+  - name: openai/gpt-3.5-turbo-instruct
+    display_name: GPT-3.5 Turbo Instruct
+    description: Similar capabilities as GPT-3 era models. Compatible with legacy Completions endpoint and not Chat Completions.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-09-18
+    tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: openai/gpt-3.5-turbo-0301
     display_name: GPT-3.5 Turbo (0301)
-    description: Sibling model of text-davinci-003 is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-03-01.
+    description: Sibling model of text-davinci-003 that is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-03-01.
     creator_organization_name: OpenAI
     access: limited
     release_date: 2023-03-01
@@ -1058,17 +1710,36 @@ models:
 
   - name: openai/gpt-3.5-turbo-0613
     display_name: GPT-3.5 Turbo (0613)
-    description: Sibling model of text-davinci-003 is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-06-13.
+    description: Sibling model of text-davinci-003 that is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-06-13.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-06-13
+    tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-3.5-turbo-1106
+    display_name: GPT-3.5 Turbo (1106)
+    description: Sibling model of text-davinci-003 that is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-11-06.
+    creator_organization_name: OpenAI
+    access: limited
+    # Actual release blog post was published on 2024-01-25:
+    # https://openai.com/blog/new-embedding-models-and-api-updates
+    release_date: 2024-01-25
+    tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-3.5-turbo-0125
+    display_name: gpt-3.5-turbo-0125
+    description: Sibling model of text-davinci-003 that is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2024-01-25.
     creator_organization_name: OpenAI
     access: limited
+    # Release blog post was published on 2024-01-25:
+    # https://openai.com/blog/new-embedding-models-and-api-updates
+    # The actual release date is unclear - it was described as "next week".
     release_date: 2023-06-13
     tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
-  # Claimed length is 16,384; we round down to 16,000 for the same reasons as explained
-  # in the openai/gpt-3.5-turbo-0613 comment
   - name: openai/gpt-3.5-turbo-16k-0613
     display_name: gpt-3.5-turbo-16k-0613
-    description: Sibling model of text-davinci-003 is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-06-13 with a longer context length of 16,384 tokens.
+    description: Sibling model of text-davinci-003 that is optimized for chat but works well for traditional completions tasks as well. Snapshot from 2023-06-13 with a longer context length of 16,384 tokens.
     creator_organization_name: OpenAI
     access: limited
     release_date: 2023-06-13
@@ -1079,7 +1750,7 @@ models:
 
   - name: openai/gpt-4-1106-preview
     display_name: GPT-4 Turbo (1106 preview)
-    description: GPT-4 Turbo (preview) is a large multimodal model that is optimized for chat but works well for traditional completions tasks. The model is cheaper and faster than the original GPT-4 model. Preview snapshot from November 6, 2023.
+    description: GPT-4 Turbo (preview) is a large multimodal model that is optimized for chat but works well for traditional completions tasks. The model is cheaper and faster than the original GPT-4 model. Preview snapshot from 2023-11-06.
     creator_organization_name: OpenAI
     access: limited
     release_date: 2023-11-06
@@ -1087,7 +1758,7 @@ models:
 
   - name: openai/gpt-4-0314
     display_name: GPT-4 (0314)
-    description: GPT-4 is a large multimodal model (currently only accepting text inputs and emitting text outputs) that is optimized for chat but works well for traditional completions tasks. Snapshot of gpt-4 from March 14th 2023.
+    description: GPT-4 is a large multimodal model (currently only accepting text inputs and emitting text outputs) that is optimized for chat but works well for traditional completions tasks. Snapshot of gpt-4 from 2023-03-14.
     creator_organization_name: OpenAI
     access: limited
     release_date: 2023-03-14
@@ -1117,6 +1788,40 @@ models:
     release_date: 2023-06-13
     tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: openai/gpt-4-0125-preview
+    display_name: GPT-4 Turbo (0125 preview)
+    description: GPT-4 Turbo (preview) is a large multimodal model that is optimized for chat but works well for traditional completions tasks. The model is cheaper and faster than the original GPT-4 model. Preview snapshot from 2023-01-25. This snapshot is intended to reduce cases of “laziness” where the model doesn’t complete a task.
+    creator_organization_name: OpenAI
+    access: limited
+    # Actual release blog post was published on 2024-01-25:
+    # https://openai.com/blog/new-embedding-models-and-api-updates
+    release_date: 2024-01-25
+    tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-4-turbo-2024-04-09
+    display_name: GPT-4 Turbo (2024-04-09)
+    description: GPT-4 Turbo (2024-04-09) is a large multimodal model that is optimized for chat but works well for traditional completions tasks. The model is cheaper and faster than the original GPT-4 model. Snapshot from 2024-04-09.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2024-04-09
+    tags: [TEXT_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-4-vision-preview
+    # According to https://platform.openai.com/docs/models/gpt-4-turbo-and-gpt-4, this model has pointed gpt-4-1106-vision-preview.
+    display_name: GPT-4V (1106 preview)
+    description: GPT-4V is a large multimodal model that accepts both text and images and is optimized for chat ([model card](https://openai.com/research/gpt-4v-system-card)).
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-11-06
+    tags: [VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  - name: openai/gpt-4-1106-vision-preview
+    display_name: GPT-4V (1106 preview)
+    description: GPT-4V is a large multimodal model that accepts both text and images and is optimized for chat ([model card](https://openai.com/research/gpt-4v-system-card)).
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2023-11-06
+    tags: [VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   ## Codex Models
   # DEPRECATED: Codex models have been shut down on March 23 2023.
@@ -1198,7 +1903,109 @@ models:
     release_date: 2022-12-15 # Blog post date
     tags: [TEXT_SIMILARITY_MODEL_TAG]
 
+  # Text-to-image models
+  - name: openai/dall-e-2
+    display_name: DALL-E 2 (3.5B)
+    description: DALL-E 2 is a encoder-decoder-based latent diffusion model trained on large-scale paired text-image datasets. The model is available via the OpenAI API ([paper](https://arxiv.org/abs/2204.06125)).
+    creator_organization_name: OpenAI
+    access: limited
+    num_parameters: 3500000000
+    release_date: 2022-04-13
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
 
+  - name: openai/dall-e-3
+    display_name: DALL-E 3
+    description: DALL-E 3 is a text-to-image generation model built natively on ChatGPT, used to prompt engineer automatically. The default style, vivid, causes the model to lean towards generating hyper-real and dramatic images. The model is available via the OpenAI API ([paper](https://cdn.openai.com/papers/dall-e-3.pdf)).
+    creator_organization_name: OpenAI
+    access: limited
+    num_parameters: 0
+    release_date: 2023-11-06
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: openai/dall-e-3-natural
+    display_name: DALL-E 3 (natural style)
+    description: DALL-E 3 is a text-to-image generation model built natively on ChatGPT, used to prompt engineer automatically. The natural style causes the model to produce more natural, less hyper-real looking images. The model is available via the OpenAI API ([paper](https://cdn.openai.com/papers/dall-e-3.pdf)).
+    creator_organization_name: OpenAI
+    access: limited
+    num_parameters: 0
+    release_date: 2023-11-06
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: openai/dall-e-3-hd
+    display_name: DALL-E 3 HD
+    description: DALL-E 3 is a text-to-image generation model built natively on ChatGPT, used to prompt engineer automatically. The HD version creates images with finer details and greater consistency across the image, but generation is slower. The default style, vivid, causes the model to lean towards generating hyper-real and dramatic images. The model is available via the OpenAI API ([paper](https://cdn.openai.com/papers/dall-e-3.pdf)).
+    creator_organization_name: OpenAI
+    access: limited
+    num_parameters: 0
+    release_date: 2023-11-06
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  - name: openai/dall-e-3-hd-natural
+    display_name: DALL-E 3 HD (natural style)
+    description: DALL-E 3 is a text-to-image generation model built natively on ChatGPT, used to prompt engineer automatically. The HD version creates images with finer details and greater consistency across the image, but generation is slower. The natural style causes the model to produce more natural, less hyper-real looking images. The model is available via the OpenAI API ([paper](https://cdn.openai.com/papers/dall-e-3.pdf)).
+    creator_organization_name: OpenAI
+    access: limited
+    num_parameters: 0
+    release_date: 2023-11-06
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
+  # Qwen
+
+  - name: qwen/qwen-7b
+    display_name: Qwen
+    description: 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. 
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2024-02-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: qwen/qwen1.5-7b
+    display_name: Qwen1.5 (7B)
+    description: 7B-parameter version of the large language model series, Qwen 1.5 (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. 
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2024-02-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: qwen/qwen1.5-14b
+    display_name: Qwen1.5 (14B)
+    description: 14B-parameter version of the large language model series, Qwen 1.5 (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. 
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2024-02-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: qwen/qwen1.5-32b
+    display_name: Qwen1.5 (32B)
+    description: 32B-parameter version of the large language model series, Qwen 1.5 (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. 
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2024-02-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: qwen/qwen1.5-72b
+    display_name: Qwen1.5 (72B)
+    description: 72B-parameter version of the large language model series, Qwen 1.5 (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. 
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2024-02-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: qwen/qwen-vl
+    display_name: Qwen-VL
+    description: Visual multimodal version of the Qwen large language model series ([paper](https://arxiv.org/abs/2308.12966)).
+    creator_organization_name: Alibaba Cloud
+    access: open
+    release_date: 2023-08-24
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  - name: qwen/qwen-vl-chat
+    display_name: Qwen-VL Chat
+    description: Chat version of Qwen-VL ([paper](https://arxiv.org/abs/2308.12966)).
+    creator_organization_name: Alibaba Cloud
+    access: open
+    release_date: 2023-08-24
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
   # Salesforce
   - name: salesforce/codegen # NOT SUPPORTED
@@ -1211,6 +2018,16 @@ models:
     tags: [] # TODO: add tags
 
 
+  # Snowflake
+  - name: snowflake/snowflake-arctic-instruct
+    display_name: Arctic Instruct
+    description: Arctic combines a 10B dense transformer model with a residual 128x3.66B MoE MLP resulting in 480B total and 17B active parameters chosen using a top-2 gating.
+    creator_organization_name: Snowflake
+    access: open
+    num_parameters: 482000000000
+    release_date: 2024-04-24
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
 
   # Stability AI
   - name: stabilityai/stablelm-base-alpha-3b
@@ -1351,6 +2168,16 @@ models:
 
 
   # Tsinghua
+
+  - name: thudm/cogview2
+    display_name: CogView2 (6B)
+    description: CogView2 is a hierarchical transformer (6B-9B-9B parameters) for text-to-image generation that supports both English and Chinese input text ([paper](https://arxiv.org/abs/2105.13290))
+    creator_organization_name: Tsinghua
+    access: open
+    num_parameters: 6000000000
+    release_date: 2022-06-15
+    tags: [TEXT_TO_IMAGE_MODEL_TAG]
+
   - name: tsinghua/glm
     display_name: GLM (130B)
     description: GLM (130B parameters) is an open bilingual (English & Chinese) bidirectional dense model that was trained using General Language Model (GLM) procedure ([paper](https://arxiv.org/pdf/2210.02414.pdf)).