Granite 4.0 H 1B

Granite 4.0 H 1B

IBM · Released Oct 28, 2025
Intelligence #688 / 699
7.2 our score
AA Index #402 / 427
2.2 Artificial Analysis
Input Price
Not priced
Output Price
Not priced
Context
Not reported

Analysis Summary

Granite 4.0 H 1B is IBM's one-billion-parameter hybrid model released in October 2025. The record includes measurements for reasoning, mathematics, coding, instruction following, long-context handling, terminal use, and tool-oriented behavior. Performance is limited across these areas, although its instruction-following result is somewhat stronger than its general reasoning and coding results. Pricing and deployment characteristics are not provided.

It could handle narrowly defined extraction or classification tasks when outputs are checked by rules. It is not suitable for autonomous agents, client-facing copy, complex SEO work, code generation, or analysis of large documents. The compact size may appeal to constrained deployments, but the available data does not establish throughput or cost benefits. Use it only after task-specific validation, and prefer a stronger small model for any workflow where errors create client or operational risk.

Assessed August 9, 2026

Editorial notes

Granite 4.0 H 1B provides limited general capability, with modest instruction-following and coding measurements. Its low reasoning and tool-use results make it unsuitable for dependable client automation.

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

DFO Verdict

Granite 4.0 H 1B provides limited general capability, with modest instruction-following and coding measurements. Its low reasoning and tool-use results make it unsuitable for dependable client automation.

#688 of 699 overall

Benchmark scores

GPQA Diamond 26.3%
HLE 5%
MMLU Pro 27.7%
AIME 2025 6.3%
SciCode 8.2%
LiveCodeBench 11.5%
τ²-Bench 19.6%
IFBench 26.2%
LCR 6.3%

Magenta = intelligence · Ink = technical/agentic · Cyan = content & long-context · Grey = community benchmarks. Data: Artificial Analysis, Hugging Face.

2.2 Intelligence Index·6.3 Math Index

How Granite 4.0 H 1B compares

Granite 4.0 H 1B ranks #402 of 427 AI models we track for overall intelligence. Granite 4.0 H 1B is currently free to use via OpenRouter.

Position in the field
Intelligence: smarter than 2% of models #688
worst in fieldmedianbest in field
Price vs frontier peers · $ per 1M tokens
Granite 4.0 H 1B $0.00 in $0.00 out
Claude Opus 5 $5.00 in $25.00 out
SpaceXAI: Grok 4.6 $2.00 in $6.00 out
Qwen: Qwen3.8 Max $2.00 in $6.00 out

Dark bar = input · light bar = output, scaled to the priciest peer.

Context window vs peers · tokens
Granite 4.0 H 1B 0

1M tokens ≈ 8 full-length novels or ~2,500 pages of business documents in a single request.

Intelligence0.8Technical0Value0Content2.2
Performance profile

Strongest on content. The pulled-in value corner is the trade-off, and if the shape matters more than the price, this is your model.

Compare shapes side-by-side →

Frequently asked questions about Granite 4.0 H 1B

How much does Granite 4.0 H 1B cost?

Granite 4.0 H 1B is currently available for free via OpenRouter.

Who created Granite 4.0 H 1B?

Granite 4.0 H 1B is developed by IBM and was released on October 28, 2025.

© 2026 Design for Online Ltd. Registered in England and Wales No. 10328553. VAT Registered. Design for Online® and Forerunner® are registered trademarks.