Working Around the Compute Ceiling: Byte-Exact Memory in Galahad Makes LLM Reading a One-Time Cost LLM Reading a One-Time Cost Paper • 2609.39358 • Published 6 days ago • 3
Working Around the Compute Ceiling: Byte-Exact Memory in Galahad Makes LLM Reading a One-Time Cost LLM Reading a One-Time Cost Paper • 2609.39358 • Published 6 days ago • 3
A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever Paper • 2607.23806 • Published Jul 26 • 8
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel Paper • 2607.14431 • Published Jul 15 • 7
A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever Paper • 2607.23806 • Published Jul 26 • 8 • 4
A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever Paper • 2607.23806 • Published Jul 26 • 8
A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever Paper • 2607.23806 • Published Jul 26 • 8
Running Agents A 12B Model Beats Fable 5 at 0 Tokens ⚡ 180/180 at 0 tokens, bit-exact, on verified work.
Running Agents A 12B Model Beats Fable 5 at 0 Tokens ⚡ 180/180 at 0 tokens, bit-exact, on verified work.
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel Paper • 2607.14431 • Published Jul 15 • 7
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel Paper • 2607.14431 • Published Jul 15 • 7 • 7
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel Paper • 2607.14431 • Published Jul 15 • 7 • 7
Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel Paper • 2607.14431 • Published Jul 15 • 7