Claude 3.5 Sonnet vs ZAYA1-74B-Preview

Pick ZAYA1-74B-Preview if you want the stronger benchmark profile. Claude 3.5 Sonnet only becomes the better choice if you would rather avoid the extra latency and token burn of a reasoning model.

ZAYA1-74B-Preview is clearly ahead on the provisional aggregate, 58 to 41. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

ZAYA1-74B-Preview's sharpest advantage is in knowledge, where it averages 64.3 against 59.4. The single biggest benchmark swing on the page is SWE-bench Verified, 49% to 53.2%.

Claude 3.5 Sonnet is also the more expensive model on tokens at $3.00 input / $15.00 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for ZAYA1-74B-Preview. That is roughly Infinityx on output cost alone. ZAYA1-74B-Preview is the reasoning model in the pair, while Claude 3.5 Sonnet is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. ZAYA1-74B-Preview gives you the larger context window at 256K, compared with 200K for Claude 3.5 Sonnet.