Qwen3.8 27B addition in words tested; GPT-4o used to compute sums in words across large numbers, with many incorrect results, per Colin Frasier and Bluesky post
Read the original at simonwillison.net→Research: Qwen3.8 27B addition in words Colin Frasier posted on Bluesky about an experiment he ran over two years ago using GPT-4o to see how well it could "compute the sum but return the answer in words"...
Original headline: "Qwen3.8 27B addition in words"
Coverage timeline
- Oct 4, 23:34 UTC Simon Willison lead source Qwen3.8 27B addition in words