AI model maker
IBM
2 models on Artificials, 1 with published test results and 2 with open weights you can download.
Where IBM stands
Its best model’s rank among all models tested on each skill area’s main test, the same tests model pages use. A longer bar means a better rank.
- OverallGranite 4.2 30B on Arena text#208 of 389
- CodingGranite 4.2 30B on Arena codingTop 49%
- MathGranite 4.1 8B on Arena math#204 of 372
- LanguageGranite 4.2 30B on Arena creative writing#254 of 387
- ReasoningGranite 4.2 30B on Arena hard prompts#195 of 389
Every test
Its best result on each published test, and that result’s rank among every model tested. Scores from different tests are never combined.
| Test | Best model | Score | Rank |
|---|---|---|---|
| Arena coding | Granite 4.2 30B | 1400 | #187 of 384 |
| Arena instruction following | Granite 4.2 30B | 1338 | #193 of 389 |
| Arena hard prompts | Granite 4.2 30B | 1363 | #195 of 389 |
| Arena text | Granite 4.2 30B | 1340 | #208 of 389 |
| Arena math | Granite 4.1 8B | 1317 | #204 of 372 |
| Arena creative writing | Granite 4.2 30B | 1266 | #254 of 387 |
| Arena WebDev | Granite 4.1 8B | 1190 | #119 of 123 |
Every model
Newest first. The price is per 1M tokens, blended three input tokens to one output token: IBM’s own listing, or else the middle price across hosts.
| Model | Released | Context | Price | Capability index |
|---|---|---|---|---|
| Granite 4.2 8B | Aug 24, 2026 | 131.1K tokens | $0.108 | Not on the index |
| Granite 4.0 Micro | Oct 2, 2025 | 131K tokens | $0.0408 | Not on the index |
Models, prices, context windows and release dates from models.dev (MIT). Test results from Arena (CC BY 4.0). A model counts here when IBM sells it, several hosts list it or a test covers it. Each figure is the source’s own; nothing is combined into a new score.