Loading timeline…
Loading timeline…
Every recorded change for this model by DeepSeek, newest first. Back to the entity page →
8 events
R1 Distill Llama 70B scores 1.52% on Terminal-Bench
R1 Distill Llama 70B scores 21.93% on τ²-bench
R1 Distill Llama 70B scores 27.55% on IFBench
R1 Distill Llama 70B scores 5.13% on Humanity's Last Exam
R1 Distill Llama 70B scores 40.2% on GPQA
R1 Distill Llama 70B scores 7.89 on Artificial Analysis Intelligence Index
R1 Distill Llama 70B: context length changed from 8192 to 128000
DeepSeek API lists R1 Distill Llama 70B at $0.8 in / $0.8 out per 1M tokens
Showing up to 400 events. Narrow by year or category, or use the changes feed for cursor-paged history. Times are UTC.