Granite 4.2 8B
Granite 4.2 8B scores 5.1 on the Artificial Analysis Intelligence Index (#60 of 65 ranked models) and costs $0.06 per million input tokens and $0.15 per million output tokens. Granite 4.2 8B is an AI model from IBM, released on Aug 31, 2026.
Intelligence
5.1
#60 of 65
Usage
—
Outside the top 20
Reading price
$0.06
for the text you give it, per 1M tokens (about 750,000 words)
Writing price
$0.15
for the text it writes, per 1M tokens (about 750,000 words)
Reads at once
~200 pages
how much text it keeps in mind (131,072 tokens)
Speed
91 per second
word pieces written; starts after 0.2s
How Granite 4.2 8B compares
| # | Model | Intelligence | Reading price | Writing price |
|---|---|---|---|---|
| 58 | LFM3-1.7B · LiquidAI | 6 | $0.08 | $0.20 |
| 59 | Nemotron 3.5 Lightning · NVIDIA | 5.5 | $0.07 | $0.20 |
| 60 | Granite 4.2 8B · IBM | 5.1 | $0.06 | $0.15 |
| 61 | Mistral Medium 3.5 · Mistral | 4.6 | $1.50 | $7.50 |
| 62 | Gemma 4 31B · Google | 4.4 | $0.20 | $0.40 |
Price and score history
Nov 17, 2026Nov 19, 2026
Nov 17, 2026Nov 19, 2026
See all 376 models View on OpenRouter ↗
Updated Nov 19, 2026. Prices and usage: OpenRouter. Intelligence: Artificial Analysis Intelligence Index. Always confirm the price with the provider before you build on it.