OpenAI · 116.8B parameters · 5.1B active
gpt-oss-120b: GPU memory requirements and which GPU to rent
OpenAI's larger open-weight model: 117 billion parameters with 5.1 billion active, experts in MXFP4 — about 61 GB as released, which fits one 80 GB or 96 GB GPU.
- Weights at FP8117 GB234 GB at 16-bit · 64 GB at 4-bit
- FP8, 8k context, 1 sequence123 GBweights + KV cache + runtime
- Suggested here2× A100$1,298 a month · dedicated
Inference
How much GPU memory gpt-oss-120b needs to serve
Total memory for one sequence at each context length, and for 16 concurrent sequences at 8k — weights, KV cache at 16-bit and runtime overhead included.
| Weights precision | Weights | 4k context | 32k context | 128k context | 8k × 16 sequences | Suggested here, 8k |
|---|---|---|---|---|---|---|
| 16-bit (BF16 / FP16) | 234 GB | 245 GB | 246 GB | 249 GB | 249 GB | 4× A100 · $2,596/mo |
| 8-bit (FP8 / INT8) | 117 GB | 123 GB | 124 GB | 128 GB | 128 GB | 2× A100 · $1,298/mo |
| 4-bit (GPTQ, AWQ, Q4) | 64 GB | 68 GB | 70 GB | 73 GB | 73 GB | 1× A100 · $649/mo |
GB are 109 bytes. Context lengths are shown for scale where the model supports them; an FP8 KV cache halves the cache part of each figure. As a mixture of experts, every expert stays resident, so memory follows the total parameter count, not the active one. Released in MXFP4: about 61 GB of weights as shipped.
Hardware
GPU servers that run gpt-oss-120b
For FP8 inference with an 8k context and one sequence: the configurations we would suggest, in order — one large card ahead of several small ones.
Fine-tuning gpt-oss-120b
Memory for a 2k-token sequence and a micro-batch of 4, with gradient checkpointing. Full fine-tuning keeps gradients and Adam states for every parameter; LoRA and QLoRA train small adapters on a frozen base.
| Method | Memory | Suggested here |
|---|---|---|
| QLoRA, 4-bit base | 78 GB | 1× RTX PRO 6000 · $749/mo |
| LoRA, 16-bit base | 254 GB | 4× A100 · $2,596/mo |
| Full fine-tune, mixed precision | 1,883 GB | Multi-node |
Architecture at a glance
Released in MXFP4: about 61 GB of weights as shipped. Figures from the model's published configuration; method on the calculator page.
gpt-oss-120b hardware: questions answered
How much VRAM does gpt-oss-120b need?
For the weights alone: about 234 GB at 16-bit, 117 GB at FP8 and 64 GB at 4-bit. Serving it at FP8 with an 8k-token context and one sequence takes about 123 GB in total once the KV cache and the runtime are counted; long contexts and many concurrent users add the KV cache on top — about 0.04 GB per 1,000 tokens at 16-bit.
Can gpt-oss-120b run on a single GPU?
Yes, at FP8 with a moderate context: one NVIDIA H200 (141 GB HBM3e) holds it, for $1,690 a month here. Longer contexts or more concurrent users may need a larger card or a second GPU.
What is the cheapest GPU setup to run gpt-oss-120b?
For FP8 inference with an 8k context and one sequence, the configuration we would suggest is 2× A100 · $1,298/mo; at 4-bit, 1× A100 · $649/mo. Suggestions favour one large card over several small ones: splitting a model across PCIe cards costs speed.
How much GPU memory do I need to fine-tune gpt-oss-120b?
With a 2k-token sequence and a micro-batch of 4: about 78 GB for QLoRA on a 4-bit base, 254 GB for LoRA on a 16-bit base, and 1,883 GB for a full fine-tune with Adam in mixed precision. Gradient checkpointing is assumed; sharding the optimizer across GPUs spreads the full fine-tune over several cards.
Run gpt-oss-120b on 2× A100 for $1,298 a month
Dedicated, single-tenant and live in under 10 minutes — with vLLM, the Hugging Face stack or Ollama preinstalled if you want. Paid in BTC, ETH or USDT, with no KYC.