Request a pilot
‹ PublicAI Index

The LLM benchmark aggregator.

Granite 4.2 30B

IBM · 30B

Strongest in Knowledge work (#93 of 178, on 1 of its 2 boards), weakest in Human preference (#177 of 342). Above par in 2 of 15 scopes. Among the models it meets almost everywhere, it finishes behind GLM-5.2 and GPT-5.6 Terra and ahead of GPT-4.1 Mini and Granite 4.2 8B.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Agents48.3#137/268
Knowledge work47.4#93/178
Tool use45.4—/81✱
Human preference51.3#177/342
Human preference51.3#177/342
Core abilities43.5—/204✱
Long context26—/0✱
Reasoning46.9—/178✱
Science44.7—/122✱
Expert reasoning36.7—/0✱
Coding49.1—/165✱
Agentic coding47.7—/157✱
Code generation50—/77✱
Knowledge48.2—/138✱
Factuality47.3—/0✱

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Granite 4.2 30B, left for the other.

Claude 3.5 Haikuwon 9 · lost 0
GPT-4o Miniwon 11 · lost 0
GPT-4.1 Nanowon 10 · lost 1
Llama 4 Scout Instructwon 8 · lost 1
Granite 4.2 3Bwon 9 · lost 3
Mistral Medium 3.5won 7 · lost 3
DeepSeek V3won 7 · lost 4
Command Awon 6 · lost 4
Mistral Large 3won 6 · lost 4
Granite 4.2 8Bwon 7 · lost 5
GPT-4.1 Miniwon 6 · lost 5
Qwen3.5 9Bwon 5 · lost 5
GPT-5.4 Nanowon 5 · lost 6
Grok 4.3won 5 · lost 6
GPT-5.4 Miniwon 5 · lost 6
Gemini 3.5 Flash Litewon 5 · lost 6
Gemini 3.1 Prowon 5 · lost 6
Gemma 4 12Bwon 4 · lost 5
Nemotron 3 Superwon 5 · lost 7
O3 Miniwon 4 · lost 6
Gemini 3.1 Flash Litewon 4 · lost 6
Nemotron 3 Ultrawon 5 · lost 8
Claude Haiku 4.5won 4 · lost 7
Inkling Smallwon 3 · lost 6
Mimo V2.5won 3 · lost 6
MiniMax M3won 5 · lost 10
Qwen3.6 35B A3Bwon 4 · lost 9
K2 Horizon 7Bwon 3 · lost 7
GPT OSS 120Bwon 3 · lost 7
GPT-5 Miniwon 3 · lost 8
Qwen3.6 Pluswon 3 · lost 8
Gemini 2.5 Prowon 3 · lost 8
Qwen3.7 Maxwon 3 · lost 8
Inklingwon 4 · lost 11
Gemma 4 31Bwon 4 · lost 11
MiniMax M2.7won 2 · lost 7
Grok 4.1 Fastwon 2 · lost 7
GPT-5.1won 2 · lost 7
Muse Sparkwon 2 · lost 7
GPT-4.1won 2 · lost 8
Kimi K2won 2 · lost 8
DeepSeek R1won 2 · lost 8
Kimi K2.5won 2 · lost 8
Gemini 3 Prowon 2 · lost 8
GLM-4.6won 2 · lost 9
Claude Sonnet 4.5won 2 · lost 9
Claude Opus 4.5won 2 · lost 9
K2 Horizon 32Bwon 2 · lost 10
Muse Glimmer-30Bwon 2 · lost 10
Grok 4.7won 1 · lost 8
GLM-4.7won 1 · lost 8
GLM-5.1won 1 · lost 8
K2 Horizon 3.7Bwon 1 · lost 9
Mimo V2.5 Prowon 1 · lost 9
Gemini 2.5 Flashwon 1 · lost 10
DeepSeek V3.2won 1 · lost 10
GPT-5.2won 1 · lost 10
Grok 4.6won 1 · lost 10
Kimi K2.6won 1 · lost 10
Grok 4.5won 1 · lost 10
GLM-5.3 Flashwon 1 · lost 10
Qwen3.8 Maxwon 1 · lost 10
Muse Spark 1.1won 1 · lost 10
Gemini 3.8 Flashwon 1 · lost 10
G9v3 39A5Bwon 1 · lost 11
DeepSeek V4 Flashwon 1 · lost 11
DeepSeek V4 Prowon 1 · lost 11
Qwen3.8 27Bwon 1 · lost 14
Claude Opus 4won 0 · lost 9
GPT-6 Lunawon 0 · lost 9
GPT-6 Solwon 0 · lost 9
GPT-6 Astrawon 0 · lost 9
Muse Spark 1.3won 0 · lost 9
Claude Opus 4.6won 0 · lost 9
Claude Opus 5.5won 0 · lost 9
Grok 4won 0 · lost 10
GPT-5won 0 · lost 10
Muse Spark 1.2won 0 · lost 10
O4 Miniwon 0 · lost 11
Claude Sonnet 4won 0 · lost 11
O3won 0 · lost 11
Claude Sonnet 4.6won 0 · lost 11
GPT-5.4won 0 · lost 11
Gemini 3.5 Flashwon 0 · lost 11
Claude Opus 4.8won 0 · lost 11
GPT-5.5won 0 · lost 11
Gemini 3.6 Flashwon 0 · lost 11
Gemini 3.7 Flashwon 0 · lost 11
Claude Fable 5.1won 0 · lost 11
Claude Opus 4.7won 0 · lost 11
Claude Fable 5won 0 · lost 11
DeepSeek V4.1 Flashwon 0 · lost 12
GLM-5.3won 0 · lost 12
GPT-5.6 Solwon 0 · lost 12
Kimi K3won 0 · lost 12
Claude Opus 5won 0 · lost 12
K2 Horizon MoVA 36B A4Bwon 0 · lost 13
K2 Horizon 375B A23Bwon 0 · lost 13
GPT-5.6 Lunawon 0 · lost 15
Claude Sonnet 5won 0 · lost 15
GPT-5.6 Terrawon 0 · lost 15
GLM-5.2won 0 · lost 15

§ 3 · Sources

Where the numbers come from

3 publications, 11 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26
LMArena Text 1341
GDPval-AA ↗1 measureread 2026-09-26
GDPval-AA 564
tau3-Banking 14.4%Terminal-Bench 2.1 26.6%SciCode 36.6%Humanity's Last Exam (without tools) 11.2%GPQA Diamond 64.4%CritPt 0.3%AA-LCR 46.7%AA-Omniscience Accuracy 10.1%AA-Omniscience Non-Hallucination 74.4%

Badge

PublicAI Index badge for Granite 4.2 30B[![PublicAI Index](https://publicai.io/model-index/badge?model=granite-4-2-30b)](https://publicai.io/model-index/m/granite-4-2-30b)

✱ Placed by a one-off publication, not a board that re-ran the model. ✱✱ marks a figure the model’s own publisher printed, which counts for less again. Scores are 0–100 on the PublicAI Index scale, 50 = the average of the models each source lists. Scores belong to their publishers.

Snapshot 2026-09-26 · the full index · JSON API