on:kimi-k2-7-code

· 6 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

@maxbittker 2026-07-17 — Kimi K3 did poorly on Runescape bench because it thinks a lot and has slow inference. Performs worse than Kimi K2.7, way ♥70
@witcheer 2026-06-22 — benchmarked Qwable-5-27b-coder on my rtx 5090. a real-trace coding SFT of Qwen3.6-27b, trained on Claude Fable-5 trace ♥37
@ultimape 2026-09-30 — @zooko ollama launch hermes --model kimi-k2.7-code:cloud And if you still run into it, there's an abliterated version i ♥3
@LinXule 2026-07-07 — this! though glm-5.2 feels more like an sonnet alternative to me (less coding & more talking and brainstorming) deepsee ♥3
@ultimape 2026-08-24 — @sidewayskoyote Honestly, my own brain works better this way as well, so that works out. I used to translate my math hom ♥1
@ultimape 2026-06-16 — @bob_obringer @techNmak I was burning $100 a day with claude. Now i'm running $20 for a kimi-k2.7:cloud with most of it ♥0