IBM

IBM Granite 4.0 Tiny Preview

Zero-eval
#3AttaQ
#3PopQA

by IBM

+
+
+
+
About

Granite 4.0 Tiny Preview was introduced as an experimental ultra-compact model, designed to demonstrate IBM's progress in efficient model development. Built to explore the boundaries of what small models can achieve for enterprise applications, it represents an early look at next-generation Granite capabilities.

+
+
+
+
Timeline
AnnouncedMay 2, 2025
ReleasedMay 2, 2025
+
+
+
+
Specifications
Training Tokens2.5T
+
+
+
+
License & Family
License
Apache 2.0
Performance Overview
Performance metrics and category breakdown

Overall Performance

12 benchmarks
Average Score
57.1%
Best Score
86.1%
High Performers (80%+)
2
+
+
+
+
All Benchmark Results for IBM Granite 4.0 Tiny Preview
Complete list of benchmark scores with detailed information
AttaQ
text
0.86
86.1%
Self-reported
HumanEval
text
0.82
82.4%
Self-reported
HumanEval+
text
0.78
78.3%
Self-reported
GSM8k
text
0.70
70.1%
Self-reported
IFEval
text
0.63
63.0%
Self-reported
MMLU
text
0.60
60.4%
Self-reported
TruthfulQA
text
0.58
58.1%
Self-reported
BIG-Bench Hard
text
0.56
55.7%
Self-reported
DROP
text
0.46
46.2%
Self-reported
AlpacaEval 2.0
text
0.35
35.2%
Self-reported
Showing 1 to 10 of 12 benchmarks
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+