Despite packing fewer shaders, NVIDIA’s RTX 4060 Ti will offer higher computational performance than its predecessor, the RTX 3070 Ti. According to a tip from @T4CFantasy, the AD106-based GPU will feature a lofty boost clock of 2.53GHz, with the AIB variants approaching the 2.7GHz mark. This makes the RTX 4060 Ti slightly faster than the RTX 3070 Ti (21.75 TFLOPs) with an FPS32 throughput of 22 TFLOPs.
RTX 4060 Ti will have a base/boost of 2310/2535 with premium aib cards up to 2685.
Edit: 4060 or 4060 Ti, Gigabyte leak changed things, it has AD106
The GeForce RTX 4060 Ti is said to pack 4,352 FP32 shaders alongside 8GB of GGDR6 memory clocked at 18Gbps. The memory bus will be 192/128-bit wide, backed by 32MB of L2 cache. The graphics card will have a TBP of 220W with a price target of around $500.
RTX 4060 Ti has a very short reference board. The PG190 still uses CEM5 connector.
8G 18Gbps GDDR6
In comparison, the GeForce RTX 3070 Ti features 6,144 shaders across 48 SMs with a peak boost clock of 1.77GHz. On the memory side, you get 8GB of GDDR6X VRAM paired with a 256-bit bus and a TBP of 290W. The RTX 3060 Ti sports 4,864 FP32 shaders (@1.66GHz) and 8GB of GDDR6 memory via a 256-bit bus. The card has a TBP of 200W and retails for $399.