ibm-granite/granite-4.2-3b vs Qwen3 4B
They cost about the same for input tokens: ibm-granite/granite-4.2-3b $0.03 per M, Qwen3 4B $0.03 per M. Qwen3 4B is 4× cheaper for output tokens: ibm-granite/granite-4.2-3b costs $0.12 per M and Qwen3 4B costs $0.03 per M. About the same: ibm-granite/granite-4.2-3b takes up to 131,072 tokens of context and Qwen3 4B up to 128,000.
| ibm-granite/granite-4.2-3b | Qwen3 4B | |
|---|---|---|
| Model id | ibm-granite/granite-4.2-3b | qwen/qwen3-4b-fp8 |
| Input | $0.03 per M | $0.03 per M |
| Output | $0.12 per M | $0.03 per M |
| Context length | 131,072 tokens | 128,000 tokens |
| Accepts | text | text |
| Produces | text | text |
| Served by | deepinfra | novita |
Prices are the providers' list prices per million tokens, as of . Both run on your own provider key through one OpenAI-compatible endpoint, with no fee on the tokens.