Alibaba merilis Qwen 3.8 minggu lalu, dan sekarang ada di DigitalOcean Serverless Inference.
The thing that got my attention is the context window, hundreds of thousands of tokens in one pass. Basically a whole codebase. So I threw an entire repo at it.
Bangku Kebun Teh Jepang - Qwen 3.8 Maks
Di Text Arena, Gemini 3.7 Flash (Tinggi) menempati posisi #9 dengan 1490 poin, peningkatan lain dari Gemini 3.6 Flash (Tinggi) (#16 -> #9).
By category, Gemini 3.7 Flash (High) also strongly improved: - #3 Creative Writing (#12 -> #3) - #4 Math (#7 -> #4) - #9 Instruction Following (
Muse-Glimmer-30B dari Meta adalah model pertama yang berjalan pada 3090 saya yang mampu melakukan satu kali pengambilan gambar ini.
Not super complicated at all, all other models couldn't one-shot and needed more guidance to get to result. Not only this, but Muse is able to one-shot almost everything i have thrown
Sudah sampai ujung arsip
Semua qwen38