Together AI's OSCAR Killed KV Cache Memory 8x — The First 2-Bit That Doesn't Collapse at 128K
中文摘要
Together AI 开源的 OSCAR 可将 KV 缓存减少 8 倍，是首个在 128K 上下文中仍能保持稳定的 2-bit 方法。
English Summary
Together AI's open-source OSCAR reduces KV cache memory 8x and is the first 2-bit method to maintain stability at 128K context.
Original Excerpt
Every 2-bit KV cache method I tried in 2025 collapsed past 32K context. Together AI’s OSCAR, open-sourced on May 25, 2026, kept Qwen3–8B… Continue reading on Towards AI »