Skip to content
ModelOpt

Model Comparison Chart

Sort by model size, VRAM minimum, or context window.

Use Cases
Llama 3.2 1B1B2GB128,000edge, rewrite, assistant-lite
Phi-3 Mini3.8B3GB4,000edge, chat, reasoning, low-latency
Llama 3.2 3B3B3GB128,000chat, summary, edge, tool-use
Mistral 7B7B5GB32,000chat, instruction-following, function-calling
DeepSeek Coder 6.7B6.7B5GB16,000coding, code-review, code-completion
Qwen 2.5 7B7B5GB32,000chat, reasoning, multilingual, json
Neural Chat 7B7B5GB32,000chat, assistant, customer-support
Yi 6B6B5GB4,096bilingual, chat, general
Llama 3.1 8B8B6GB128,000chat, general, instruction-following, rag
Qwen 2.5 Coder 7B7.6B6GB32,768coding, refactor, test-generation, agents
Gemma 2 9B9B7GB8,192chat, analysis, summarization
Mistral NeMo 12B12B8GB128,000chat, coding, reasoning, long-context
Solar 10.7B10.7B8GB4,096chat, single-turn, general
Phi-3 Medium14B10GB128,000reasoning, chat, analysis
Qwen 2.5 14B14B10GB32,000reasoning, agent, multilingual, json
Codestral 22B22B16GB32,768coding, fim, repository-analysis
Gemma 2 27B27B20GB8,192analysis, reasoning, enterprise-chat
DeepSeek Coder 33B33B22GB16,000coding, repository-analysis, test-generation
Yi 34B34B24GB4,096analysis, bilingual, knowledge
Llama 3.1 70B70B40GB128,000advanced-reasoning, agent, analysis, multilingual