Model creator

IBM

Every IBM model in the BenchSift catalogue, with its scores, ranks, prices and capabilities as the sources publish them.

View in the catalogue

Models: 14
Open weights: 14; 14 of 14 models
Latest release: Aug 25, 2026; Granite 4.2 30B · First dated model: Apr 16, 2025
Median price: $0.096; $0.052 to $0.282 per million tokens, across 6 priced models

Best scores and ranks

The creator's best model on each measure, and its rank among every measured model in the catalogue.

Intelligence

Granite 4.1 30B

7.4

Coding

Granite 4.2 30B

29.9

Math

Granite 4.0 H Small

13.7

Agentic

Granite 4.2 30B

3.5

Output speed

Granite 4.1 8B

126 t/s

Lowest price

Granite 4.2 3B

$0.052

OpenRouter ranking

Granite 4.2 8B

#242

Capabilities

Number of models that declare each capability. Tool calling and structured outputs come from OpenRouter.

Reasoning3 of 14

Open weights14 of 14

Tool calling1 of 14

Structured outputs1 of 14

Input

Text14

Output

Text14

Overview

Model types

Text14

Releases by year

20266
20258

Models

Reasoning levels are folded into one entry. The rank is the whole catalogue's: sorting does not change it.

Granite 4.1 30B
Text · Open weights · Apr 29, 2026
Apr 29, 20267.4#567 of 782——
Granite 4.1 8B
Text · Open weights · Apr 29, 2026 · $0.063 · 126 t/s
Apr 29, 20266.6#622 of 782$0.063126 t/s
Granite 4.0 H Small
Text · Open weights · Sep 22, 2025 · $0.107 · 14.4 t/s
Sep 22, 20256.0#660 of 782$0.10714.4 t/s
Granite 4.1 3B
Text · Open weights · Apr 29, 2026
Apr 29, 20265.9#670 of 782——
Granite 4.2 30B
Text · Reasoning · Open weights · Aug 25, 2026 · $0.282 · 41.4 t/s
Aug 25, 20265.5#698 of 782$0.28241.4 t/s
Granite 4.0 H 1B
Text · Open weights · Oct 28, 2025
Oct 28, 20255.2#728 of 782——
Granite 4.2 8B
Text · Reasoning · Open weights · Aug 25, 2026 · $0.107 · 102 t/s
Aug 25, 20265.1#739 of 782$0.107102 t/s
Granite 4.0 Micro
Text · Open weights · Sep 22, 2025
Sep 22, 20255.1#739 of 782——
Granite 4.0 1B
Text · Open weights · Oct 28, 2025
Oct 28, 20255.0#746 of 782——
Granite 3.3 8B (Non-reasoning)
Text · Open weights · Apr 16, 2025 · $0.085 · 45.4 t/s
Apr 16, 20254.9#753 of 782$0.08545.4 t/s
Granite 4.0 350M
Text · Open weights · Oct 28, 2025
Oct 28, 20254.8#757 of 782——
Granite 4.0 H 350M
Text · Open weights · Oct 28, 2025
Oct 28, 20254.8#757 of 782——
Granite 4.2 3B
Text · Reasoning · Open weights · Aug 25, 2026 · $0.052 · 59.2 t/s
Aug 25, 20264.7#774 of 782$0.05259.2 t/s
Granite-4.0-H-Micro
Text · Open weights · Oct 2, 2025
Oct 2, 2025———

Where these figures come from

Each catalogue model is tied to its creator by the identity Artificial Analysis publishes, or OpenRouter's and Models.dev's for models it does not track. Counts fold the reasoning levels of one model together. Ranks compare each measure with every measured model in the catalogue, configurations included, as model pages do. The price is the published blended price (3 input tokens to 1 output token).

Figures computed by BenchSift from the catalogue: Artificial Analysis, OpenRouter, Models.dev and Hugging Face. A value no model carries is not shown.

Data retrieved on Nov 12, 2026.