gpt-oss-120b — VRAM requirements & what it costs to run
Bottom line: at q4_k_m with 8k context you need ≈ 73.2 GiB; smallest single accelerator in our catalog: A100 80GB.
Quantization table (GiB)
| quant | weights | total @8k | need @8k | smallest fit |
|---|---|---|---|---|
| native | 60.77 | 67.68 | 67.7 | A100 80GB |
| bf16 | 217.61 | 240.21 | 240.2 | needs multi-GPU / offload |
| fp8 | 108.81 | 120.52 | 120.5 | H200 141GB |
| q8_0 | 115.61 | 128.0 | 128.0 | H200 141GB |
| q6_k | 89.22 | 98.98 | 99.0 | H200 141GB |
| q5_k_m | 77.12 | 85.67 | 85.7 | H200 141GB |
| q4_k_m | 65.83 | 73.25 | 73.2 | A100 80GB |
| q3_k_m | 53.18 | 59.34 | 59.3 | A100 80GB |
| q2_k | 45.56 | 50.96 | 51.0 | A100 80GB |
Dataset: modelvram.com open dataset (CC BY 4.0 / MIT), retrieved 2026-10-20. "total @8k" already includes KV cache + runtime overhead per dataset methodology; we extrapolate KV linearly by context (formula).
KV-cache scaling at q4_k_m (total GiB)
| context | total |
|---|---|
| 4k | 73.1 |
| 8k | 73.2 |
| 16k | 73.5 |
| 32k | 74.1 |
| 128k | 77.5 |
Cost: rent vs buy vs API
Cheapest listed rental that fits: RunPod A100 80GB @ $1.59/h — source: modelvram.com public table (self-stated check 2026-09-29, retrieved 2026-10-20). Partial re-verify 2026-10-20: Lambda A100 40GB $1.99/h CONFIRMED on lambda.ai/pricing (fetched directly); Lambda H100 rows ambiguous on-page ($3.99 vs $4.09 configs) → not adopted; RunPod & Vast pricing pages are JS-rendered, no extractable static figures — verify rates visually before spending
Cross-check on record for this family: vLLM official recipes, retrieved 2026-10-20: FP8 ≈ 200 GB, NVFP4 ≈ 170 GB; Thunder Compute blog, 2026-08-12: native checkpoint ≈ 175 GB, 3-bit GGUF ≈ 131 GB. Datasets differ by quantisation bookkeeping — see method.
Running gpt-oss-120b with Ollama
Typical q4_k_m GGUF download ≈ 65.83 GiB (weights, dataset figure); expect ≥ 73.2 GiB resident at 8k context. Find tags: ollama.com/search (we do not mirror tag lists — sizes on Ollama are their numbers, ours are dataset-sourced).
FAQ
How much VRAM does gpt-oss-120b need?
At q4_k_m with an 8k context it needs about 73.2 GiB including KV cache and runtime overhead; the smallest card in our catalog that fits is A100 80GB. Numbers come from the modelvram.com open dataset (CC BY 4.0, retrieved 2026-10-20) and our published formula (https://llmbudgetlab.com/method/).
Can I run gpt-oss-120b on a Mac?
Use the calculator: macOS caps GPU-usable memory well below your spec-sheet unified RAM. We count 2/3 of unified (3/4 above 36GB) as a deliberately conservative rule of ours, not an Apple constant — compare required GiB against usable memory, then confirm on your own machine with sysctl iogpu.wired_limit_mb (why).
Is renting a GPU cheaper than buying one for gpt-oss-120b?
Rental from $1.59/h (RunPod) — plug both into the calculator's break-even field; buy price and electricity are yours to enter because we refuse to guess market prices.
**gpt-oss-120b** needs ~73.2 GiB VRAM at q4_k_m (8k context) → minimum A100 80GB. Check & cost-compare: [LLM Budget Lab](https://llmbudgetlab.com/models/gpt-oss-120b/)or grab the SVG badge →
More: builds under $1000 · under $1500 · used GPUs & break-even · RTX 4090 · RTX 5090