一周开源生态盘点：V4.1 Flash 在 Terminal-Bench 4.0 上以 1/20 成本胜 GPT-5.6

RANA @LeaderX_btc原推发布于 2026.10.25

在 X 上打开原推

Look at what shipped across the open ecosystem in a single week: • Cost & Speed: DeepSeek-V4.1-Flash beats GPT-5.6 Sol on Terminal-Bench 4.0 at 20x lower cost, while Liquid AI’s LFM2.5 runs private on-device agents with just 2B active parameters per token. • Production ROI: Coinbase fine-tuned an open Qwen 3.5 9B model to cut fraudulent transactions by 30%. • Agent-Native Tooling: Keel built an MIT-licensed Git platform where agents merge by testing "belief collisions" rather than line diffs.