IBM: Granite 4.1 8B

IBM: Granite 4.1 8B

ibm-granite · Released Apr 30, 2026 Professional
Intelligence #14 / 590
82.0 Our Score
Speed #100 / 279
127.6 tokens / sec
Input #157 / 592
$0.050 per 1M tokens
Output #142 / 592
$0.100 per 1M tokens
Context #245 / 592
131,072 tokens

Analysis Summary

IBM Granite 4.1 8B is a small enterprise-focused model from IBM with tool use and function calling support, priced at $0.05/$0.10 per million tokens. Its benchmark scores are modest: reasoning and coding capability are limited, and agentic performance is below the competitive threshold for complex workflows. Instruction following and long-context reliability are also constrained.

For businesses, the model's main appeal is its very low cost and tool use capability, which makes it viable for high-volume, simple automation tasks where answer quality requirements are low. It is not suited to coding assistance, complex document analysis, or customer-facing content where accuracy matters.

At this price point, it competes with other small open-weight models. Teams in the IBM ecosystem may find it a convenient fit for lightweight classification, routing, or structured extraction tasks, but should not rely on it for anything requiring strong reasoning.

Assessed June 30, 2026

Editorial notes

IBM Granite 4.1 8B is a compact, low-cost model with tool use and function calling support, but limited reasoning and coding benchmarks restrict it to simple, high-volume tasks.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence1.4Technical2.6Value8Content2.9
Intelligence 1.4/10
Technical 2.6/10
Content 2.9/10
Value 8/10

How IBM: Granite 4.1 8B compares

IBM: Granite 4.1 8B ranks #295 of 385 AI models we track for overall intelligence, #127 of 139 for coding, #151 of 293 for agentic tasks. Its 131K-token context window is larger than 59% of the models we list. At $0.05 per million input tokens it is cheaper than 73% of comparable models.

About IBM: Granite 4.1 8B

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..

8B Parameters

Capabilities

Tool Use Function Calling

Performance Indices

Source: Artificial Analysis

6.7 Intelligence Index
9.5 Coding Index
27.8 Agentic Index

Benchmark Scores

Intelligence

GPQA Diamond 43.3% Graduate-level scientific reasoning
HLE 3.8% Humanity's Last Exam
SciCode 21.8% Scientific computing

Technical

τ²-Bench 27.8% Conversational agent benchmark

Content

IFBench 38.6% Instruction following
LCR 12% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does IBM: Granite 4.1 8B stack up?

Compare side-by-side with other professional models.

Compare Models

Model Information

OpenRouter ID ibm-granite/granite-4.1-8b
Provideribm-granite
Release Date April 30, 2026
Context Length131,072 tokens
Max Completion131,072 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.05 $0.000050
Output $0.10 $0.000100

Live Performance

Live endpoint metrics, refreshed every 30 minutes.

100%
Avg Uptime
177ms
Best Latency (TTFT)
117 tok/s
Best Throughput
1/1
Active Endpoints
Available via: WandB

Leaderboard Categories

Frequently asked questions about IBM: Granite 4.1 8B

How much does IBM: Granite 4.1 8B cost?

IBM: Granite 4.1 8B costs $0.05 per million input tokens and $0.10 per million output tokens.

What is the context window of IBM: Granite 4.1 8B?

IBM: Granite 4.1 8B has a context window of 131,072 tokens (131K).

Is IBM: Granite 4.1 8B good for coding?

On our coding benchmark index, IBM: Granite 4.1 8B ranks #127 of 139 models, placing it in the broader range of the field for code generation and debugging.

What can IBM: Granite 4.1 8B do?

IBM: Granite 4.1 8B supports tool use and function calling.

Who created IBM: Granite 4.1 8B?

IBM: Granite 4.1 8B is developed by IBM and was released on April 30, 2026.