RTX 4090 vs RTX 3090 for local LLMs

Computed with the open FitLLM engine — accurate per-layer KV-cache modeling, not a naive estimate. Updated 2026-06.

FitLLM compares fit — what loads in memory — computed from official config.json. These are a floor, not a guarantee; speed and power are not estimated.

The two cards

	RTX 4090	RTX 3090
VRAM	24 GB	24 GB
Memory bandwidth (speed, not estimated)	1008 GB/s	936 GB/s

What each runs (~4-bit, max context that fits)

Model	RTX 4090	RTX 3090
Qwen 3.6 35B-A3B	❌ won't fit · 24.8/24 GB	❌ won't fit · 24.8/24 GB
Qwen 3.6 27B	✅ up to 34K · 20.2/24 GB	✅ up to 34K · 20.2/24 GB
Gemma 4 31b	⚠️ up to 3K · 23.5/24 GB	⚠️ up to 3K · 23.5/24 GB
Gemma 4 26b A4B	✅ up to 83K · 18.9/24 GB	✅ up to 83K · 18.9/24 GB
Gemma 4 12b	✅ up to 262K · 10.4/24 GB	✅ up to 262K · 10.4/24 GB

For model fit, the RTX 4090 and RTX 3090 are interchangeable — same VRAM, same verdicts. The faster card wins on memory bandwidth and power, not capacity. FitLLM only claims fit, not speed.

Bottom line

For local LLMs these two are interchangeable on capacity — choose on speed, price and power, not on what fits.

All numbers are computed by the open-source fitllm-engine (MIT) from official model config.json values — reproduce or audit them yourself. Estimates; real usage varies with runtime (llama.cpp / MLX / Ollama), driver and display. Found a mismatch? Report it. · FitLLM home