Skip to content

A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.

Revamp Ollama models list (#7595)

vincent d warmerdamaa2dbcf6dafa
diff --git a/packages/llm-info/data/models.yml b/packages/llm-info/data/models.yml
index 9902947e29..ceecf88f92 100644
--- a/packages/llm-info/data/models.yml
+++ b/packages/llm-info/data/models.yml
@@ -8,7 +8,7 @@
providers: [anthropic]
roles: [chat, edit]
thinking: true
-
+
- name: Claude 4.5 Sonnet
model: claude-sonnet-4-5
description: Latest Sonnet model with enhanced capabilities and performance improvements
@@ -232,7 +232,6 @@
roles: [chat, edit]
thinking: false
-
# Bedrock
- name: Claude 4.5 Sonnet (Bedrock)
@@ -293,146 +292,27 @@
# Ollama Models
-- name: Llama 3.1 8B
- model: llama3.1:8b
- description: Meta's efficient 8B parameter model with excellent chat and coding capabilities
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Llama 3.1 70B
- model: llama3.1:70b
- description: Large-scale Meta model with superior reasoning and complex task performance
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Llama 3.3 70B
- model: llama3.3:70b
- description: Latest Llama model with enhanced capabilities and performance improvements
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Mistral 7B
- model: mistral:7b
- description: Efficient 7B parameter model with strong performance for general tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Mixtral 8x7B
- model: mixtral:8x7b
- description: Mixture of experts model with 8x7B parameters for enhanced performance
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Mixtral 8x22B
- model: mixtral:8x22b
- description: Large mixture of experts model with superior reasoning capabilities
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Gemma 2 9B
- model: gemma2:9b
- description: Google's efficient 9B parameter model optimized for performance and safety
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Gemma 2 27B
- model: gemma2:27b
- description: Larger Gemma model with enhanced capabilities for complex tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Qwen 2.5 7B
- model: qwen2.5:7b
- description: Alibaba's multilingual model with strong performance across diverse tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Qwen 2.5 14B
- model: qwen2.5:14b
- description: Mid-size Qwen model balancing performance and efficiency
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Qwen 2.5 32B
- model: qwen2.5:32b
- description: Large Qwen model with excellent reasoning and multilingual capabilities
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Qwen 2.5 72B
- model: qwen2.5:72b
- description: Flagship Qwen model with top-tier performance for complex reasoning tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: DeepSeek R1 8B
- model: deepseek-r1:8b
- description: Efficient reasoning model with strong analytical capabilities
+- name: gpt-oss 20b
+ model: gpt-oss:20b
+ description: OpenAI’s open-weight models designed for powerful reasoning, agentic tasks, and versatile developer use cases.
providers: [ollama]
roles: [chat, edit]
thinking: true
-- name: DeepSeek R1 32B
- model: deepseek-r1:32b
- description: Advanced reasoning model with enhanced problem-solving abilities
+- name: Qwen3 8b
+ model: qwen3:8b
+ description: Qwen3 is the latest generation of large language models in Qwen series.
providers: [ollama]
roles: [chat, edit]
thinking: true
-- name: DeepSeek R1 70B
- model: deepseek-r1:70b
- description: Flagship reasoning model with superior analytical and thinking capabilities
+- name: Gemma3 12b
+ model: gemma3:12b
+ description: Lightweight gemma models trained by Google, trained to run on a single GPU.
providers: [ollama]
roles: [chat, edit]
thinking: true
-- name: CodeLlama 34B
- model: codellama:34b
- description: Meta's specialized code generation model with strong programming capabilities
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: DeepSeek Coder 33B
- model: deepseek-coder:33b
- description: Specialized coding model with excellent programming and software development skills
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Qwen 2.5 Coder 7B
- model: qwen2.5-coder:7b
- description: Code-optimized version of Qwen 2.5 for programming tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: Phi-4 14B
- model: phi4:14b
- description: Microsoft's compact yet capable model optimized for reasoning tasks
- providers: [ollama]
- roles: [chat, edit]
- thinking: false
-
-- name: TinyLlama 1.1B
- model: tinyllama:1.1b
- description: Ultra-lightweight model for resource-constrained environments
- providers: [ollama]
- roles: [chat]
- thinking: false
-
# Weights & Biases Models
- name: Meta Llama 3.1 8B (W&B)
@@ -490,4 +370,3 @@
providers: [wandb]
roles: [chat, edit]
thinking: true
-