model roundup

Gemini 3

2 items · started 2026-04-16 · closed 2026-04-19

Holy moly — Qwen3-35B-A3B-UD-IQ2_M just surpassed Gemini 3 Flash at coding, running on my RX 9070 XT at 99 tok/sec (www.reddit.com)

8 7w qwen gemini

So I ran a small personal test giving both models the same coding tasks. For A* Pathfinding, Qwen absolutely crushed Gemini 3 Flash — both in code quality and overall thoughtfulness.
Llamaindex releases Parsebench (www.reddit.com)

+2 7w qwen gemini

https://preview.redd.it/c0ns26pf3mvg1.png?width=1920&format=png&auto=webp&s=4b6ac114c2e0395684ac0ba79e591d71ccca2fe3 ParseBench lets you test the accuracy of different parsers using your own documents. Ran this across Gemini 3 flash, Qwen…