Why BQN Wins
Summary
The article presents a benchmark comparing BQN's performance for an odd binary function, highlighting its advantages with a live demonstration.
View Cached Full Text
Cached at: 08/18/26, 02:13 AM
codereport/max-odd-binary
Source: https://github.com/codereport/max-odd-binary
🥳 LIVE BENCHMARK: HERE 🥳
Similar Articles
@no_stp_on_snek: someone will wave the card at me: the 9B crushes its base on the coding benchmarks (SWE-bench 69 vs 53). true. but on m…
A commentator discusses the performance of a 9B model on coding benchmarks, noting that while it beats its base on SWE-bench (69 vs 53), the advantage narrows on behavioral and long-horizon tests, suggesting limited gains outside benchmark distributions.
I benchmarked PrismML's 1-bit Bonsai-8B against IBM's Granite on CPU tool calling. The 1-bit model won, but only with grammar-constrained decoding
An independent benchmark of PrismML's 1-bit Bonsai-8B against IBM's Granite and other models on CPU tool calling shows that with grammar-constrained decoding, Bonsai-8B achieves a 92% pass rate, outperforming larger models, but fails without constraints. Granite is the best raw model at 72%.
Spent two weeks on a kernel that benchmarked 29x faster. End to end it's maybe 6-10%, and it's not even wired in yet.
The author optimized a matmul kernel for BitNet's ternary models on CPU, achieving 29x speedup in isolation, but found that the model is memory-bound, resulting in only 6-10% end-to-end gain. The inference engine is available as open-source.
ProgramBench Vetted: Reverse Engineering from a Runnable Binary
ProgramBench Vetted is a benchmark with 50 tasks that evaluate AI agents on reconstructing programs from runnable binaries, designed to study long-context coordination and enable dense reward for reinforcement learning.
Bias Compounds, Variance Washes Out
This article demonstrates that using stochastic rounding for BF16 optimizer state can match FP32 performance because unbiased errors cancel over time, whereas round-to-nearest stalls due to compounding bias. An experiment with an MLP shows BF16+SR achieves similar loss to FP32 while using less memory.