IBM
Granite 4.2 8B Instruct开源
IBM Granite 4.2 8B lightweight enterprise foundation model with high throughput and low cost.
Intelligence
49.2 #338
AAQI 43.0
Output speed
115 t/s
TTFT 200 ms · ITL 8.7 ms/tok
Blended price (3:1)
$0.22 /1M
In $0.15/M · Out $0.45/M
Context
128K
≈ 320 A4 pages · max out 8192
Quality Domain Radar
6-Axis Frontier Evaluation vs Category Average & Flagship
Granite 4.2 8B Instruct
Claude Mythos Preview (#1)
Average (68)
Radar axes are approximated from category scores where direct benchmark measurements are missing — see 评分方法.
API deployment
Compare hosts & latency
Throughput, TTFT and cache pricing across Groq, Cerebras, Together, DeepInfra, Fireworks, Bedrock and Azure.
GroqCerebrasTogether AIDeepInfraFireworks AIAWS BedrockAzure AI
Groq
Cerebras
Together AI
DeepInfra
Fireworks AI
AWS Bedrock
Azure AI各榜单得分
分类能力得分
基准测试分项
模型详细参数
发布日期
2026年7月20日
参数量
8 billion
开源协议
Open Weights
上下文窗口
128,000 tokens
输入价格
$0.15/M
输出价格
$0.45/M
输出速度
115 t/s
首Token延迟
200 ms
Estimated monthly bill
Estimated monthly spend
$0.00
Calculating savings vs GPT-4o...
Cite & embed
GitHub Markdown badge
[](https://llmpodium.com/models/granite-4-2-8b)BibTeX
@misc{llmpodium2026granite428b,
title = {Benchmark Analysis and Intelligence Evaluation of Granite 4.2 8B Instruct},
author = {LLMPodium Research},
year = {2026},
url = {https://llmpodium.com/models/granite-4-2-8b}
}More from IBM
替代模型推荐
基于能力、定价和基准测试得分与 Granite 4.2 8B Instruct 最接近的替代模型。