impulse

IBM

Open-weight labwww.ibm.com/granite
Models
28
not retired
Lineup
1
last 18 months
Hosts
5
Retiring
0
next 12 months

Current lineup

General-purpose models released in the last 18 months, newest first. Prices are the lab's own, per million tokens.

ModelIn / out per 1M
$0.06 / $0.15
via W&B Inference

Images, audio, search and more

ModelIn / out per 1M
Moderation
$0.10 / $0.10
Moderation
$0.20 / $0.20
Transcription
$0.00 / $4.90
via Replicate
Transcription
$0.03 / $0.25
via Replicate

Older models

Superseded or undated, but still served somewhere.

Show 23 older models
ModelIn / out per 1M
Chat
$0.05 / $0.25
via Replicate
Chat
$0.20 / $0.20
Chat
$0.60 / $0.60
Chat
$0.60 / $0.60
Chat
$0.10 / $0.50
via Replicate
Chat
$0.20 / $0.20
Chat
$0.03 / $0.25
via Replicate
Chat
-
Chat
$0.03 / $0.25
via Replicate
Chat
$0.03 / $0.25
via Replicate
Chat
$0.03 / $0.25
via Replicate
Chat
-
Chat
$0.0636 / $0.265
Chat
$0.017 / $0.112
via Cloudflare Workers AI
Chat
$0.06 / $0.25
via Replicate
Chat
-
Chat
$0.05 / $0.25
via Replicate
Chat
$0.38 / $0.38
Chat
$0.38 / $0.38
Chat
$0.38 / $0.38
Chat
$0.10 / $0.10
Chat
$0.00 / $0.25
via Replicate
Chat
$0.06 / $0.25
via Replicate

Where to run IBM models

Hosts by how many of its models they serve.