RTX 5080 vs RTX 4080 SUPER for local LLMs

Computed with the open FitLLM engine — accurate per-layer KV-cache modeling, not a naive estimate. Updated 2026-07-16.

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 5080	RTX 4080 SUPER
VRAM	16 GB	16 GB
Memory bandwidth (speed, not estimated)	960 GB/s	736 GB/s

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

Model	RTX 5080	RTX 4080 SUPER
Hy3	❌ won't fit · 196/16 GB	❌ won't fit · 196/16 GB
GLM-5.2	❌ won't fit · 484/16 GB	❌ won't fit · 484/16 GB
GLM-4.7-Flash	❌ won't fit · 21.9/16 GB	❌ won't fit · 21.9/16 GB
gpt-oss-20b	⚠️ won't fit · 15.9/16 GB	⚠️ won't fit · 15.9/16 GB
gpt-oss-120b	❌ won't fit · 77.2/16 GB	❌ won't fit · 77.2/16 GB
Qwen 3.6 35B-A3B	❌ won't fit · 24.8/16 GB	❌ won't fit · 24.8/16 GB
Qwen 3.6 27B	❌ won't fit · 20.2/16 GB	❌ won't fit · 20.2/16 GB
Qwen-AgentWorld-35B-A3B	❌ won't fit · 24.6/16 GB	❌ won't fit · 24.6/16 GB
Gemma 4 31b	❌ won't fit · 23.5/16 GB	❌ won't fit · 23.5/16 GB
Gemma 4 26b A4B	❌ won't fit · 18.9/16 GB	❌ won't fit · 18.9/16 GB
Gemma 4 12b	✅ up to 110K · 10.4/16 GB	✅ up to 110K · 10.4/16 GB
Llama-3.1-8B-Instruct	✅ up to 48K · 8.5/16 GB	✅ up to 48K · 8.5/16 GB
Llama-3.2-3B-Instruct	✅ up to 73K · 5.3/16 GB	✅ up to 73K · 5.3/16 GB
MiniCPM5-1B	✅ up to 131K · 3.2/16 GB	✅ up to 131K · 3.2/16 GB

For model fit, the RTX 5080 and RTX 4080 SUPER 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