IBM: Granite 4.1 8B

Text input Text output Unavailable
Author's Description

Granite 4.1 8B is a dense, decoder-only 8-billion-parameter language model from IBM, part of the Granite 4.1 family. It supports a 131K-token context window and is designed for enterprise tasks...

Key Specifications
Cost
$$
Context
131K
Parameters
8B
Released
Apr 30, 2026
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Tool Choice Tools Response Format Temperature Max Tokens Structured Outputs Presence Penalty Stop Top P Frequency Penalty Seed
Features

This model supports the following features:

Structured Outputs Tools Response Format
Performance Summary

IBM Granite 4.1 8B demonstrates competitive response times, ranking in the 57th percentile for speed across various benchmarks. It consistently offers among the most competitive pricing, placing in the 88th percentile. The model exhibits exceptional reliability with a 99% success rate, indicating minimal technical failures. In terms of performance across categories, Granite 4.1 8B shows strong capabilities in Hallucinations (98.0% accuracy, 73rd percentile) and Ethics (99.0% accuracy, 59th percentile). Its Email Classification performance is a standout, achieving 95.0% accuracy and ranking #1 in speed, demonstrating near-perfect accuracy at the highest speed. However, the model shows notable weaknesses in Instruction Following (15.0% accuracy, 20th percentile), Reasoning (42.0% accuracy, 22nd percentile), and Mathematics (68.0% accuracy, 25th percentile), where its accuracy falls into lower percentiles. General Knowledge and Coding performance are moderate, at 91.5% and 78.0% accuracy respectively, but with lower percentile rankings.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.05
Completion $0.1
Input Cache Read $0.05

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
WandB
WandB | ibm-granite/granite-4.1-8b-20260429 131K $0.05 / 1M tokens $0.1 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by ibm-granite