IBM: Granite 4.1 8B
The IBM Granite 4.1 8B model is designed for extensive text processing with a context length of up to 131072 tokens and can handle up to 131072 output tokens. It supports input in text format and integrates tools, though it does not include advanced reasoning or structured output capabilities. For those prioritizing cost-effectiveness and moderate benchmark performance across two independent tests, the IBM Granite 4.1 8B stands out with its blended score of 10.3; however, given a low benchmark coverage of only 2, users should consider this in light of the pricing-$0.05 per million input tokens and $0.1 per million output tokens-which may be higher than some alternatives.
Benchmark results
Independent, published benchmarks. Blended score 10.3 across 2 benchmarks, last refreshed 2026-08-13. How scoring works →
| Benchmark | Measures | Score |
|---|---|---|
| AI Index | broad capability composite | 10.6 |
| AI Index Coding | software engineering tasks | 15.7 |
- Model ID
- ibm-granite/granite-4.1-8b
- Vendor
- ibm-granite
- Released
- April 2026
- Tokenizer
- Other
- Input Modalities
- text
- Output Modalities
- text
- Max Output
- 131,072 tokens
- Tool Calling
- ✓ supported
- Structured Output
- ✓ supported
- Reasoning Mode
- not supported
- Vision
- text only
- Audio
- no
- Moderated
- no
What it costs in practice
Computed from the current $0.05/M input and $0.10/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | under $0.01 |
| Classify 1,000 customer emails | 500k in / 50k out | $0.03 |
| A month of a busy support chatbot | 5M in / 2M out | $0.45 |
Similar models
Quick answers
- How much does IBM: Granite 4.1 8B cost?
- $0.05 per million input tokens and $0.10 per million output tokens.
- What is IBM: Granite 4.1 8B's context window?
- 131,072 tokens, roughly 196 pages of text in a single request.
- Does IBM: Granite 4.1 8B support tool calling?
- It supports tool calling, structured output.
- Can IBM: Granite 4.1 8B process images?
- No, it is text-only on the input side.