Models · Compare
Granite 4.0 Small vs Claude Sonnet 4.5
Rows highlighted in warm gray are where the models differ. Numbers carry their as-of date and primary source.
Specs
| Field | A: Granite 4.0 Small | B: Claude Sonnet 4.5 |
|---|---|---|
| Released | 2025-10-02 | 2025-09-29 |
| Developer | IBM | Anthropic |
| Openness | Open | Proprietary |
| License | Apache-2.0 | Proprietary |
| OSI-approved | yes | no |
| Data released | no | no |
| Training code | no | no |
| Architecture | hybrid-mamba-transformer-moe | unknown |
| Total params | 32B | — |
| Active params | 9B | — |
| Experts | — | — |
| Context window | 131K | — |
| Attention | hybrid-mamba2-transformer | unknown |
| Position enc. | unknown | unknown |
| Pretraining tokens | — | — |
| Post-training | sft, rlhf | rlhf, constitutional |
| Training hardware | — | — |
| $/M input | — | $3.00 |
| $/M output | — | $15.00 |
| Output tok/sec | — | 48.8 |
Benchmarks
Missing scores render as not reported; never inferred. Bold highlights the leader per benchmark.
General reasoning
| MMLU-Pro | — | 86.0 2026-05-21 |
| GPQA-Diamond | — | 72.7 2026-05-21 |
Code
| SWE-Bench Verified | — | 77.2 2025-09-29 |
| LiveCodeBench | — | 59.0 2026-05-21 |
Math
| AIME 2025 | — | 37.0 2026-05-21 |
Context · A
IBM's first hybrid Mamba/Transformer release, October 29 2025. Granite-4.0-H-Small is a 32B MoE activating 9B per token, with Mamba-2 and conventional transformer blocks layered 9:1. Shipped alongside H-Tiny (7B / 1B) and H-Micro (3B dense), all base + instruct, all Apache 2.0.
Context · B
September 29 2025 incremental update of the Sonnet line at the same $3 / $15 price point as Sonnet 4. Shipped alongside Claude Code checkpoints, a native VS Code extension, context editing and memory tools for long-running agents, and the Claude Agent SDK. OSWorld score of 61.4% positioned it as the strongest computer-use model from Anthropic at launch.