The Tema_Q Series is an LLM model series developed to achieve both unrestricted dialogue (uncensored) and high practical reasoning performance.
| Model Name | Parameters | Status | Features & Achievements |
|---|---|---|---|
| Tema_Q-R1.0 | 9B | Unreleased | Prototype model |
| Tema_Q-R2.0-Code | 9B | Under Evaluation | An uncensored model specialized for code writing |
| Tema_Q-R2.5 | 7B | Unreleased | |
| Tema_Q-R3.1 | 9B | Excellent | First release. An uncensored model boasting one of the highest NatInt scores for an indie LLM |
| Tema_Q-R4.1 | 10B | Unreleased | |
| Tema_Q-R4.2 | 10B | Excellent | Achieved the highest score in the under-22B category on the UGI Leaderboard |
| Tema_Q-R5.0 | 9B | Unreleased | |
| Tema_Q-R6.0 | 9B | Under Evaluation | Feedback model for the development of next-generation versions |
| Tema_Q-R7.0 | 12B | Highly Rated | A well-balanced 12B parameter uncensored model |
| Tema_Q-R-4B | 4B | Highly Rated | A well-balanced 4B parameter uncensored model |
| Tema_Q-R8.1 | 10B | Unreleased | |
| Tema_Q-R8.5 | 10B | Unreleased | |
| Tema_Q-R9.0 | 10B | Unreleased | |
| Tema_Q-R10.1 | 12B | Unreleased | |
| Tema_Q-R10.2 | 9B | Unreleased | |
| Tema_Q-R10.5 | 12B | Unreleased | |
| Tema_Q-R11.5 | 9B | Unreleased | |
| Tema_Q-R12.2 | 9B | Under Evaluation | A Japanese-specialized model trained on a massive 400B-parameter class dataset.While he has high intelligence due to his enhancements, his writing ability is slightly low |
| Tema_Q-R13 | 12B | Unreleased | |
| Tema_Q-R14.3 | 12B | Unreleased | |
| Tema_Q-R15.6 | 9B | Highly Rated | High-precision Japanese model that recorded a score of 4.10 on ELYZA-tasks-100 |
| Tema_Q-X | 10B | Excellent | The flagship model that combines an ultimate uncensored environment with high accuracy |
| Tema_Q-X-Thinking | 14B | Under Evaluation | An uncensored inference model based on DeepSeek. It is an uncensored model that is low in censorship but capable of advanced inference |
| Tema_Q-X-P1 | 12B | Unreleased | Prototype model for 12B model |
| Tema_Q-X-4B | 4B | Excellent | Highly uncensored Instruct model |
| Tema_Q-X-4B-Thinking | 4B | Excellent | Highly uncensored Thinking model |
| Tema_Q-X2 | 30B | Excellent | It's incredibly smart and runs four times faster than traditional methods |
| Tema_Q-X2-Thinking | 9B | Excellent | Very clever model |
| Tema_Q-X2-4B-Thinking | 4B | Under Evaluation | |
| Tema_Q-X3 | 80B | Excellent | |
| Tema_Q-X3-Thinking | 9B | Excellent |