RTX 3080 for local LLMs — what fits in 10 GB & the real cost
Rent a RTX 3080 from $0.14/hr ($0 markup, as of 2026-07-02). 235 of the 445 models in our catalog fit its 10 GB — plus the honest own-vs-rent-vs-API math, electricity and amortization included.
Three honest ways to run AI on a RTX 3080
Own it: buy the card and pay electricity + amortization — local is real money, not "$0".
Rent it: from $0.14/GPU-hr on your own vendor account (as of 2026-07-02, $0 markup — we never resell compute).
Skip it: run a comparable open model through your own API key, paying per million tokens (prices as of 2026-07-06).
Rent a RTX 3080 — real vendor rates, $0 markup
Vast.ai: $0.14/GPU-hr — vendor's own price as of 2026-07-02; you pay them directly.
What fits in 10 GB
235 of the 445 models in the Spanvero catalog fit on one RTX 3080 at a sensible quant (context capped at 16k for the estimate). Most capable first:
Mellum2 12B A2.5B Thinking (JetBrains, 12.1B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.20/1M tokens (size estimate).
Mellum2 12B A2.5B Base (JetBrains, 12.1B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.20/1M tokens (size estimate).
pythia 12b (EleutherAI, 12B) — ~10 GB at Q4_K_M; own-key API alternative ≈ $0.20/1M tokens (size estimate).
Falcon3-10B Instruct (TII, 10B) — ~10 GB at Q4_K_M; own-key API alternative ≈ $0.18/1M tokens (size estimate).
gemma 2 9b it (unsloth, 9.2B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
Gemma 2 9B Instruct (Google, 9B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.06/1M tokens (last-known).
NVIDIA Nemotron Nano 9B v2 (nvidia, 8.9B) — ~10 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
NVIDIA Nemotron Nano 9B v2 Japanese (nvidia, 8.9B) — ~10 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
granite 4.1 8b (ibm-granite, 8.8B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.08/1M tokens.
internlm3 8b instruct (internlm, 8.8B) — ~8 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
llm jp 4 8b thinking (llm-jp, 8.6B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
Nemotron Labs Diffusion 8B Base (nvidia, 8.5B) — ~7 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
Nemotron Labs Diffusion 8B (nvidia, 8.5B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
GDN primed HQwen3 8B Instruct (amazon, 8.5B) — ~9 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
LFM2.5 8B A1B (LiquidAI, 8.5B) — ~7 GB at Q4_K_M; own-key API alternative ≈ $0.17/1M tokens (size estimate).
Used NVIDIA RTX 3080 10GB (used market): street price $285.00 (as of 2026-07-08; BestValueGPU eBay used-price tracker, Jun 2026 (~$284.97, reading dated Jun 23 2026)) → $0.2603/day amortized over 3 years.
Electricity ≈ $0.2232/1M tokens at 320 W and $0.19/kWh (EIA Electric Power Monthly Table 5.6.A — U.S. residential average, Apr 2026, as of 2026-07-02), assuming ~75 tok/s (8B-class model at Q4).
vs renting: owning wins above ~3.3 active hours/day at the cheapest logged rate ($0.14/hr).
vs API: Llama 3.1 8B Instruct at $0.03/1M (as of 2026-07-06) is cheaper than the electricity alone — owning never breaks even at any volume.
As an Amazon Associate we earn from qualifying purchases. Each link opens an Amazon search for the category — not an endorsement of any specific product — and Cynosure LLC may earn a commission at no extra cost to you.
A short email of real AI price moves, straight from the daily log — no hype. We're collecting the list now; the first issue goes out when it opens. Unsubscribe with one click.