Falcon H1 Tiny R 0.6B
Falcon H1 Tiny R 0.6B is a downloadable reasoning model with a hybrid transformer and Mamba architecture. It is part of TII's experiment in training small models directly on reasoning data.
How this checkpoint was trained
The 0.6B release has both pre-GRPO and post-GRPO checkpoints. The main repository is the post-GRPO version. This makes it possible to study what reinforcement learning changed without comparing unrelated base models.Falcon H1 Tiny reasoning study
Use a fixed reasoning budget
Practical guidance. Use a fixed generation budget when comparing the two checkpoints. Record answer accuracy, unfinished responses and repeated text separately. Choose the post-training stage explicitly in the model identifier so a reproduction does not silently compare different variants.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | tiiuae/Falcon-H1-Tiny-R-0.6B |
| License | Falcon-LLM LicenseFalcon H1 Tiny R 0.6B model card |
| Nominal size | 600M parametersFalcon H1 Tiny R 0.6B model card |
Related models
- Falcon H1 Tiny R 90M
Technology Innovation Institute (TII) · Falcon H1 Tiny
Falcon H1 Tiny R 90M is a downloadable reasoning model with a hybrid transformer and Mamba architecture.
Launch date not confirmedDownloadable - Falcon H1R 7B
Technology Innovation Institute (TII) · Falcon H1R
Falcon H1R 7B is a reasoning-focused adaptation of Falcon-H1-7B-Base.
Launch date not confirmedDownloadable