Stats
Actions
Tags
From nvidia-models
Fan a prompt to 3 NVIDIA models and synthesize agreements/divergences
How this command is triggered — by the user, by Claude, or both
Slash command
/nvidia-models:nvidia-compare <prompt to send to all three models>The summary Claude sees in its command listing — used to decide when to auto-load this command
Call `mcp__nvidia-models__nvidia_compare` with this prompt: > $ARGUMENTS Default `models` to `["llama-4-maverick", "deepseek-v4-pro", "qwen3-next-80b"]` unless the user specified others. Use `max_tokens: 1024`. After the tool returns: 1. Show a short table with one row per model and a 1-2 sentence distillation of its answer. 2. Highlight where the three answers agree (consensus) and where they diverge (note which model said what). 3. Pick the most useful answer and explain why — or, if they're complementary, synthesize a combined response. Do NOT dump the raw concatenated output unless...
Call mcp__nvidia-models__nvidia_compare with this prompt:
$ARGUMENTS
Default models to ["llama-4-maverick", "deepseek-v4-pro", "qwen3-next-80b"] unless the user specified others. Use max_tokens: 1024.
After the tool returns:
Do NOT dump the raw concatenated output unless the user explicitly asks to see it.
npx claudepluginhub najemwehbe/nvidia-models-plugin --plugin nvidia-models