Zyphra
/

ZR1-1.5B

@@ -17,7 +17,7 @@ pipeline_tag: text-generation
 ZR1-1.5B is a small reasoning model trained extensively on both verified coding and mathematics problems with reinforcement learning. The model outperforms Llama-3.1-70B-Instruct on hard coding tasks while achieving strong scores on math evaluations and a 37.91% pass@1 accuracy on GPQA-Diamond with just 1.5B parameters.
-![ZR1-1.5B LiveBench evaluation results: the model is very token efficient](zr1_livebench.png)
 ## Data

 ZR1-1.5B is a small reasoning model trained extensively on both verified coding and mathematics problems with reinforcement learning. The model outperforms Llama-3.1-70B-Instruct on hard coding tasks while achieving strong scores on math evaluations and a 37.91% pass@1 accuracy on GPQA-Diamond with just 1.5B parameters.
+![ZR1-1.5B LiveBench evaluation results on LiveBench with greedy sampling: the model is very token efficient](zr1_livebench_greedy.png)
 ## Data