ModelPriceWatch.com
Last scan 2026-08-09 Models tracked 198 Providers 30 Cheapest paid Granite 4.0 Micro $0.017/Mtok in Every price links to its source

IBM

Commercial provider · 5 models tracked · Founded 1911

IBM's open-weight Granite 4 family for enterprise workloads — Granite 4 H Large down to Granite 4.0 Micro at $0.017/1M input tokens.

IBM pricing at a glanceAugust 2026 · $ per 1M tokens

IBM API pricing (August 2026): 5 current models range from $0.017 to $0.300 per 1M input tokens and $0.112 to $1.20 per 1M output tokens. The cheapest paid model is Granite 4.0 Micro at $0.017/1M input; the priciest is Granite 4 H Large at $0.300/1M input / $1.20 output. Every price links to IBM's official pricing page and refreshes twice daily.

Models5
Input range$0.017–$0.300
Output range$0.112–$1.20
Cached-input tiers0

Today's IBM prices

5 models · cheapest blended first

Sorted by blended cost (cheapest first). Prices per 1M tokens, August 2026 — every price links to IBM's official pricing page.

Current IBM model prices per 1M tokens, sorted by blended cost
Model Blended* Input Output Cached in Relative cost Context Status
Granite 4.0 Micro
$0.041
$0.017 $0.112
128K Current
Granite Embedding 278M Multilingual
$0.106
$0.106 $—
Current
Granite 4 H Small
$0.107
$0.060 $0.250
128K Current
Granite 4 H Medium
$0.262
$0.150 $0.600
128K Current
Granite 4 H Large
$0.525
$0.300 $1.20
128K Current
* Blended = (3×input + 1×output) ÷ 4 $/1M tokens · cheap · mid · expensive MODELPRICEWATCH.COM · 2026-08-09

Quick stats

Models tracked
5
Type
commercial
Founded
1911
Cheapest model
Granite 4.0 Micro
Cheapest blended
$0.041/M

Open-weight models

5 open-weight models available from this provider.

Granite 4 H Large
$0.300/M in · 128K ctx
Granite 4 H Medium
$0.150/M in · 128K ctx
Granite 4 H Small
$0.060/M in · 128K ctx
Granite 4.0 Micro
$0.017/M in · 128K ctx
Granite Embedding 278M Multilingual
$0.106/M in · — ctx

Try IBM

Sign up and start building with IBM models.

Get started
Run it yourself

Self-host IBM's open-weight models

These models ship with open weights, so you can serve them yourself on rented GPUs instead of paying per-token API prices — often cheaper at scale.

Partner links — we may earn a commission if you sign up, at no cost to you. We only list platforms we'd recommend regardless.