---
last_updated: 2026-11-22
source_url: https://opper.ai/ibm/granite-4-2-8b
title: "Granite 4.2 8B by IBM"
description: "Granite 4.2 8B is IBM's 8B dense reasoning model with built-in thinking that can be switched off or set to low effort, made for tool calling, code and multilingual dialog in 12 tested languages."
---

# Granite 4.2 8B

By [IBM](https://opper.ai/ibm).

Released 25 Aug 2026 · [Announcement](https://huggingface.co/ibm-granite/granite-4.2-8b) · [All IBM releases](https://opper.ai/model-releases/ibm)

Granite 4.2 8B is the mid-size reasoning model of IBM's Granite 4.2 family, released on August 25, 2026 and post-trained from Granite 4.1 8B Base. It is a dense decoder-only transformer that reasons step by step inside think tags before it answers; thinking is on by default and can be switched off per request or reduced to a low-effort mode when latency matters more than depth.

IBM trained it with supervised fine-tuning and several rounds of reinforcement learning, including agentic phases on software engineering, terminal use and web search with trajectories of up to 200 tool-calling turns, so it plans which tools to call and why. It natively supports a 128K-token context, with a long-context extension to 512K, and is tested in English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch and Chinese.

The weights are released under Apache 2.0, alongside 3B and 30B siblings in the same family.

## Available routes

Granite 4.2 8B runs on 1 route through the Opper gateway. Prices are per 1M tokens. Its routes and prices, live from the catalogue, are public JSON with no API key needed at https://api.opper.ai/v3/models?q=deepinfra%2Fibm-granite%2Fgranite-4.2-8b.

| Provider | Model id | Region | Zero data retention | Training | Input | Output | Cache read |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [DeepInfra](https://opper.ai/provider/deepinfra) | `deepinfra/ibm-granite/granite-4.2-8b` | US (United States) | Yes | No | $0.06 | $0.25 | $0.01 |

## Key info

- Model id: `deepinfra/ibm-granite/granite-4.2-8b`
- Input: Text
- Output: Text
- Features: Reasoning, Tools, Structured output
- Context window: 131K
- Max output: n/a
- Input price: $0.06 per 1M tokens ($0.01 cache read)
- Output price: $0.25 per 1M tokens
- Released: 25 Aug 2026
- US residency available
- Zero data retention on pay-as-you-go
- No training by default

## Data handling per route

Each route hosting Granite 4.2 8B has its own privacy posture, residency, and GDPR terms. Postures are maintained by Opper with a last-verification timestamp.

### DeepInfra — United States

Zero data retention: nothing is logged, held for abuse monitoring or used for training. No training on customer data. US; SCCs.

- Zero data retention: Yes. Nothing is logged, held for abuse monitoring or used for training.
- Training: No training on customer data.
- Logging: None
- Abuse monitoring: No classifier
- Transfer mechanism: SCCs

Verified by Opper on 2026-10-19.

## Best for

Where Granite 4.2 8B ranks among the models on Opper with a score for each use case (0 to 100).

- [Legal](https://opper.ai/models?use=legal): 14.6 (Legal tasks and reasoning)
- [Finance](https://opper.ai/models?use=finance): 14.8 (Finance and accounting)
- [Coding](https://opper.ai/models?use=coding): 0.0 (FrontierCode 2.0 and Terminal-Bench 4.0)
- [Agents](https://opper.ai/models?use=agents): 1.1 (Terminal-Bench 4.0 and AutomationBench)

## Benchmarks

Independent benchmark scores via [Artificial Analysis](https://artificialanalysis.ai) ([view on AA](https://artificialanalysis.ai/models/granite-4-2-8b)). Indices and domain scores are out of 100, Omniscience runs from −100 to 100, Briefcase is an Elo rating, and the other evaluation scores are the share of tasks solved.

- Global rank: #75 of 520 LLMs on the [LLM leaderboard](https://opper.ai/llm-leaderboard)
- Tier: Efficient
- Output speed: 136 tok/s
- Intelligence Index: 5.1

### Coding

- FrontierCode 2.0: 0%
- Terminal-Bench 4.0: 0%
- SciCode: 32%

### Agentic & knowledge work

- AutomationBench: 2%
- GDPval: 0%
- GDP.pdf: 1%
- Briefcase (Elo): 312

### Science & reasoning

- Terminal-Bench Science: 0%
- CritPt (hard): 0%
- HLE Diamond: 4%
- Humanity's Last Exam: 10%
- GPQA Diamond: 63%
- Long-context reasoning: 45%

### Knowledge

- Omniscience: -16.7

### By domain

- Finance & accounting: 14.8
- Legal: 14.6
- Engineering: 14.6
- Economics: 14.6
- Strategy & operations: 6.9

## Get started

Call Granite 4.2 8B through the Opper gateway with one API key. Opper is drop-in compatible with the OpenAI, Anthropic and Google AI SDKs.

Granite 4.2 8B is a premium model on Opper. Sign up needs no credit card: you get an API key straight away and the free models work in the playground and the API. Add a card to use premium models, pay-as-you-go with no minimum.

### Set it up with your agent

Paste this into a coding agent like Claude Code, Cursor or Codex and it will wire up Opper for you:

```text
Use curl to download, read and follow: https://skills.opper.ai Call Granite 4.2 8B
```

### Or call it directly

OpenAI:

```javascript
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.OPPER_API_KEY,
  baseURL: "https://api.opper.ai/v3/compat",
});

const completion = await client.chat.completions.create({
  model: "deepinfra/ibm-granite/granite-4.2-8b",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(completion.choices[0].message.content);
```

Anthropic:

```javascript
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  apiKey: process.env.OPPER_API_KEY,
  baseURL: "https://api.opper.ai/v3/compat",
});

const message = await client.messages.create({
  model: "deepinfra/ibm-granite/granite-4.2-8b",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello" }],
});
console.log(message.content);
```

Gemini:

```javascript
import { GoogleGenAI } from "@google/genai";

const ai = new GoogleGenAI({
  apiKey: process.env.OPPER_API_KEY,
  httpOptions: { baseUrl: "https://api.opper.ai/v3/compat" },
});

const response = await ai.models.generateContent({
  model: "deepinfra/ibm-granite/granite-4.2-8b",
  contents: "Hello",
});
console.log(response.text);
```

cURL:

```bash
curl https://api.opper.ai/v3/compat/chat/completions \
  -H "Authorization: Bearer $OPPER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepinfra/ibm-granite/granite-4.2-8b",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'
```

## Use Granite 4.2 8B in your apps

Pick your tool to see how to run it on Granite 4.2 8B through Opper.

- [Claude Code](https://opper.ai/apps/claude-code/granite-4-2-8b)
- [Claude Cowork](https://opper.ai/apps/claude-cowork/granite-4-2-8b)
- [Codex](https://opper.ai/apps/codex/granite-4-2-8b)
- [DeepSeek Harness](https://opper.ai/apps/deepseek-harness/granite-4-2-8b)
- [GitHub Copilot CLI](https://opper.ai/apps/github-copilot-cli/granite-4-2-8b)
- [Aider](https://opper.ai/apps/aider/granite-4-2-8b)
- [Goose](https://opper.ai/apps/goose/granite-4-2-8b)
- [Droid](https://opper.ai/apps/droid/granite-4-2-8b)
- [OpenCode](https://opper.ai/apps/opencode/granite-4-2-8b)
- [Pi](https://opper.ai/apps/pi/granite-4-2-8b)
- [OpenClaw](https://opper.ai/apps/openclaw/granite-4-2-8b)
- [Hermes](https://opper.ai/apps/hermes/granite-4-2-8b)
- [Copilot for Obsidian](https://opper.ai/apps/obsidian-copilot/granite-4-2-8b)
- [OmniRoute](https://opper.ai/apps/omniroute/granite-4-2-8b)
- [Cursor](https://opper.ai/apps/cursor/granite-4-2-8b)
- [Cline](https://opper.ai/apps/cline/granite-4-2-8b)
- [Continue](https://opper.ai/apps/continue/granite-4-2-8b)
- [Kilo Code](https://opper.ai/apps/kilo-code/granite-4-2-8b)
- [GitHub Copilot](https://opper.ai/apps/github-copilot/granite-4-2-8b)
- [Zed](https://opper.ai/apps/zed/granite-4-2-8b)
- [JetBrains AI Assistant](https://opper.ai/apps/jetbrains-ai-assistant/granite-4-2-8b)
- [Voidleap Code](https://opper.ai/apps/voidleap-code/granite-4-2-8b)
- [n8n](https://opper.ai/apps/n8n/granite-4-2-8b)
- [LibreChat](https://opper.ai/apps/librechat/granite-4-2-8b)
- [Open WebUI](https://opper.ai/apps/open-webui/granite-4-2-8b)
- [See all apps](https://opper.ai/apps)

## Other models from IBM

- [Opper Docling OCR](https://opper.ai/ibm/docling-latest)
