4 ms·
There's a chap called Bijian Bowen who does very quick agentic coding challenges for new models (very soon after release!) mainly for toy games or websites. He
by verytrivial 16d ago
There's a chap called Bijian Bowen who does very quick agentic coding challenges for new models (very soon after release!) mainly for toy games or websites. He just did one for this model and included a comparison with the base model Qwen 3.8 which shows the "near-lossless" claim should be taken with a grain of salt. It is an interesting model if you are GPU starved and want local, but you might have trouble finding things it is good at.
- Aurornis 15d ago> which shows the "near-lossless" claim should be taken with a grain of salt I agree. I don’t know how they get such good results on these benchmarks because using them gives a very different experience. They’re kind of cool for doing short free form outputs in memory constrained systems, but I don’t think they’re useful as coding agents.
- UrineSqueegee 15d agothese are very saturated benchmarks
- qingcharles 15d agoThe only two I trust are Bijan and simonw.