GPU · 149.7 TFLOPS dense BF16

A40 fine-tuning cost estimates

All sixteen scenarios on the A40: GPU-hours, wall-clock on 8 GPUs and the cost range across 3 providers, at 35% MFU. Prices read 12 November 2026. Estimates, not quotes.

The figure

Where 149.7 TFLOPS comes from

“Peak BF16 Tensor TFLOPS with FP32 Accumulate 149.7 · 299.4*”

nvidia.com

Scenarios

Cost by scenario

Scenario	GPU-hours	8 GPUs	Lowest	Median	Highest
7B LoRA fine-tune on 100 million tokens	15	1.9 h	$8.76	$25	$39
7B LoRA fine-tune on 1 billion tokens	148	19 h	$88	$254	$395
7B full fine-tune on 100 million tokens	22	2.8 h	$13	$38	$59
7B full fine-tune on 1 billion tokens	223	28 h	$131	$381	$592
13B LoRA fine-tune on 100 million tokens	28	3.4 h	$16	$47	$73
13B LoRA fine-tune on 1 billion tokens	276	34 h	$163	$472	$733
13B full fine-tune on 100 million tokens	41	5.2 h	$24	$71	$110
13B full fine-tune on 1 billion tokens	414	2.2 days	$244	$708	$1,100
34B LoRA fine-tune on 100 million tokens	72	9.0 h	$43	$123	$192
34B LoRA fine-tune on 1 billion tokens	721	3.8 days	$425	$1,234	$1,918
34B full fine-tune on 100 million tokens	108	14 h	$64	$185	$288
34B full fine-tune on 1 billion tokens	1,082	5.6 days	$638	$1,852	$2,877
70B LoRA fine-tune on 100 million tokens	148	19 h	$88	$254	$395
70B LoRA fine-tune on 1 billion tokens	1,484	7.7 days	$876	$2,541	$3,949
70B full fine-tune on 100 million tokens	223	28 h	$131	$381	$592
70B full fine-tune on 1 billion tokens	2,227	11.6 days	$1,314	$3,812	$5,923

Prices used

A40 price by provider

Provider	$/GPU-hr	Configuration	Source
RunPod	$0.59	1×	price page
Vultr	$1.71	1×	price page
Runcrateours, live	$2.66	1×	price page

Each provider's lowest listed on-demand configuration for this GPU. Marketplace figures are left out.

Questions

A40: common questions

What is the lowest cost scenario on the A40?

7B LoRA fine-tune on 100 million tokens: about $8.76 on RunPod, 15 GPU-hours.

What throughput is assumed?

149.7 TFLOPS dense BF16 as printed by the maker, times 35% MFU, so about 52 TFLOPS of useful work.

How many providers list it?

3 with a current provider-set on-demand price. Runcrate is one of them at $2.66 live, marked as ours.

Disclosure

Fine-Tune Cost Lab is published by the Runcrate team, who also run Runcrate, an AI cloud. Runcrate is listed in the table for A40, marked "ours, live". It is sorted by price like every other row, with no pinned position. Other providers' prices are read from their own public pricing pages. We have no paid placement.