Skip to content

Cost reference

What it costs to train a 3B model from scratch

A compute-optimal run for a 3B-parameter model, 60B tokens at 20 per parameter, takes about 758 H100-hours, around $3,329 at $4.39 an hour. Training on 3,000B tokens (3T tokens, 1,000 per parameter) raises that to about $166,456. Thousands of dollars at the Chinchilla ratio; a model you'd ship, trained on trillions of tokens, is a six-figure run.

By Will Schulz · Checked 15 November 2026 · Prices as of November 2026

Compute-optimal: 60B tokens

GPUGPU-hoursCostTime on 8Time on 64Time on 512GPUs to hold state
RTX 5090209.5 TFLOPS · $1.39/hr3,580$4,97619 days2 days7.0 hours2
A100 80GB312 TFLOPS · $2.19/hr2,404$5,26413 days38 hours4.7 hours1
H100 SXM989 TFLOPS · $4.39/hr758$3,3294 days12 hours1.5 hours1
H200989 TFLOPS · $5.29/hr758$4,0124 days12 hours1.5 hours1
B2002,250 TFLOPS · $7.99/hr333$2,66342 hours5.2 hours39 minutes1

Longer training: 3,000B tokens

GPUGPU-hoursCostTime on 8Time on 64Time on 512GPUs to hold state
RTX 5090209.5 TFLOPS · $1.39/hr178,998$248,8072.6 years3.8 months15 days2
A100 80GB312 TFLOPS · $2.19/hr120,192$263,22120.6 months2.6 months10 days1
H100 SXM989 TFLOPS · $4.39/hr37,917$166,4566.5 months25 days3 days1
H200989 TFLOPS · $5.29/hr37,917$200,5816.5 months25 days3 days1
B2002,250 TFLOPS · $7.99/hr16,667$133,1672.9 months11 days33 hours1

Compute = 6 × 3e9 parameters × tokens, at 40% utilisation of dense BF16 peak, with perfect scaling across GPUs. Memory for weights, gradients and Adam state is 48 GB before activations.

What real runs of this size reported

RunSize and dataReported
Llama 3.2 3Bup to 9T tokens460K H100-hours

Published GPU-hours for Llama 3.1 came in 1.2–2.9× above the 6ND estimate; budget for that gap, plus experiments, failed runs and data work, before you treat these numbers as a quote.

Change the assumptions

Pre-filled for 3B parameters at 60B tokens on H100s.

Estimated cost

$3,329

H100 SXM-hours

758

Time on 8 GPUs

4 days

Compute

1.1e21 FLOPs

Full training needs about 48 GB for weights, gradients and optimiser state before activations, so at least 1 H100 SXM just to hold the model. Fine-tuning with QLoRA instead needs about 2 GB for 4-bit weights, which fits on a single 32 GB card.

Frequently asked questions

How much does it cost to train a 3B model?
About $3,329 on H100s at $4.39/hour for a compute-optimal run of 60B tokens (6 × parameters × tokens at 40% utilisation). At 3,000B tokens (3T tokens, 1,000 per parameter) it is about $166,456. Treat both as floors: published runs used 1.2–2.9× more GPU-hours than the formula.
How long does it take to train a 3B model?
About 4 days on one 8-GPU H100 node for 60B tokens, assuming perfect scaling, or 12 hours on 64 GPUs.
How much GPU memory does training a 3B model need?
About 48 GB for weights, gradients and Adam state (16 bytes per parameter) before activations, so at least 1 × 80 GB GPU with the state sharded.
Which GPU is cheapest for training a 3B model?
At on-demand November 2026 rates, the B200 ($7.99/hour) gives the lowest compute cost, about $2,663, by our estimate. Peak FLOPS aren't delivered performance, so benchmark your own code before a large run.

Fine-tuning instead

Open models of this size already carry far more pretraining than most budgets allow. Adapting one costs a tiny fraction:

The method in depth is in the free Academy course The Economics of Model Training and the article What it really costs to train an LLM; terms are defined in the glossary under 6ND, Chinchilla scaling and MFU.

Sources

  1. [1]Kaplan et al., 2020. Scaling Laws for Neural Language Models (arXiv:2001.08361)
  2. [2]Hoffmann et al., 2022. Training Compute-Optimal Large Language Models (arXiv:2203.15556)
  3. [3]Meta, 2024. The Llama 3 Herd of Models (arXiv:2407.21783)
  4. [4]Rajbhandari et al., 2019. ZeRO: Memory Optimizations Toward Training Trillion Parameter Models (arXiv:1910.02054)
  5. [5]Runpod. GPU pricing (updated 12 November 2026)
  6. [6]Meta. Llama 3.2 model card

You probably don't need to pretrain.

Kojen's prompt tuning and model search run on hosted models against your own test cases, and text LoRA trains on real GPUs (for accounts with GPU training enabled) when you need weights to change.

Start training free