Hello NVIDIA team,
I recently discovered the NIM API catalog on build.nvidia.com and am
genuinely impressed by how many open-weight frontier models are available.
I’d like to start a personal exploration / learning project where I run
the same prompts (reasoning puzzles, math word problems, multilingual
text) across several models side by side — Kimi K2.6, GLM-5.1, DeepSeek
V4 Pro, Qwen 3.5 397B, MiniMax M2.7 are on my shortlist.
Before I start hitting the rate limit during parallel calls, could you
raise my account from 40 RPM to 200 RPM? I haven’t made significant
traffic yet — I’d rather start with proper headroom than learn the limit
by getting 429s.
Hobby use only, no commercial scraping. Thank you!