IBM

IBM Granite 4.2 8B

Name: IBM Granite 4.2 8B
Author: IBM

IBM Granite 4.2 8B is a IBM model with 131K context, Not verified in / Not verified out per million tokens, last verified October 2, 2026. Open weights: yes. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.

Granite 4.2 8B is IBM's dense text reasoning checkpoint for smaller deployments, coding and tool use.

Recorded profile verification October 2, 2026Individual provider fields may change

AI model specification details
Specification	IBM Granite 4.2 8B
Provider	IBM
Tier	Budget
Context window	131K
Max output	Not verifiedUnverified
Input / 1M tokens	Not verifiedUnverified
Output / 1M tokens	Not verifiedUnverified
Weights	Open
Parameters	8B dense
Reasoning levels	Not verifiedUnverified
Modalities	text
License	Apache 2.0
Released	August 25, 2026

Pricing tiers: Apache 2.0 downloadable checkpoint with self-hosting compute costs. The exact config gives 131,072 native positions; IBM describes extension to 512K, which is not the default config. No current first-party hosted token rate or universal hard output cap is verified.

Publisher evidence

Published benchmark results

Each result keeps its source date visible. Dates refer to source publication or capture unless the source explicitly states when it measured the result. A missing benchmark is not treated as a zero.

Published benchmark results for IBM Granite 4.2 8B
Benchmark	Score	Source date	Source
Artificial Analysis Intelligence IndexGranite 4.2 8B	11	2026-10-01	Artificial Analysis · View source

Best for

Self-hosted text workflows
Coding and tool use
Multilingual workloads

Watch out

Thinking is enabled by default; enable_thinking=False disables it, and low_effort=True shortens enabled thinking. These are checkpoint template controls, not a low/medium/high API menu. Previous reasoning is stripped by default. IBM recommends temperature=1.0/top_p=0.95; example max_new_tokens values are not hard output caps. Check runtime memory requirements for this checkpoint size. Scores from other Granite sizes and publisher estimates do not transfer.

Source receipts

Catalog figures for this model were checked against the following sources.

IBM exact 8B card: August 25 release, license and modes (accessed 2026-10-02)
IBM exact 8B native context config (accessed 2026-10-02)
Artificial Analysis October 1 exact variant table; estimates withheld (accessed 2026-10-02)

Common questions

IBM Granite 4.2 8B

Answered from the verified figures on this page rather than general guidance.

What is IBM Granite 4.2 8B's context window?

IBM Granite 4.2 8B accepts about 131K tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.

What is IBM Granite 4.2 8B best for?

IBM Granite 4.2 8B is a budget tier from IBM. It suits self-hosted text workflows, coding and tool use, multilingual workloads. Thinking is enabled by default; enable_thinking=False disables it, and low_effort=True shortens enabled thinking. These are checkpoint template controls, not a low/medium/high API menu. Previous reasoning is stripped by default. IBM recommends temperature=1.0/top_p=0.95; example max_new_tokens values are not hard output caps. Check runtime memory requirements for this checkpoint size. Scores from other Granite sizes and publisher estimates do not transfer.

Can I self-host IBM Granite 4.2 8B?

IBM Granite 4.2 8B publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.

Model comparison catalog Benchmark library