● FitLLM — fit receipt · Q4_K_M · 8K tokens ctx · KV F16 · engine 2.8.1

gpt-oss-120b on RTX 3090

WON'T FIT ✗

weights	KV cache	overhead	reserve	total	short by
66.7 GB	0.3 GB	8.3 GB	2.0 GB	77.2 / 24 GB	53.2 GB

→ A100 80GB would fit.

Replay: npx fitllm-engine "gpt-oss-120b" --gpu "RTX 3090" · JSON · interactive

— embed: ![fits](https://img.shields.io/endpoint?url=https%3A%2F%2Ffitllm.run%2Fapi%2Fbadge%3Fmodel%3Dgpt-oss-120b%26gpu%3DRTX%2B3090%26quant%3DQ4_K_M%26ctx%3D8192%26kv%3D16)

Computed by the open fitllm-engine (MIT) from official config.json values — estimates; runtime varies. Ran it for real? Challenge this prediction — measured reports calibrate the engine for everyone.