FIGFLEX: A Mixed-Precision FP-INT GEMM Accelerator Co-Designed with Low-Bit LLM Quantization

No. 237 of 245 · arXiv · Faraz Tahmasebi, Haocheng Xu, Andrew Trauger et al. · published 23 Oct 2026, 04:00 UTC · first published 21 Oct 2026

AI & machine learning
Hardware & devices
Paper

2 judges. Scored by two of the three judges. The third reads only the day’s strongest candidates. How far stories are checked

FIGFLEX is a hardware-software co-design for running low-bit LLMs, where weights are stored as integers and multiplied against floating-point activations. Its accelerator uses a precision-flexible radix-4 datapath so that throughput rises as weight precision falls, including an INT3 mode that uses otherwise idle hardware.

Read the original at arxiv.org. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Score

0.55total under rubric v2.0

Significance0.55
Reader interest0.64
Audience breadth0.42
Practical relevance0.54
Substantiated0.63
Technical depth0.82
New information0.87
Counts against: Promotional0.09
Counts against: Overclaim0.36

Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).

Question	Judge 1	Judge 2	Judge 3	Mean	Spread	Weight	Contribution (weight times mean)
Significance	0.65	0.45	—	0.55	0.20	+0.25	+0.138
Reader interest	0.67	0.60	—	0.64	0.07	+0.20	+0.127
Audience breadth	0.43	0.40	—	0.42	0.03	+0.10	+0.042
Practical relevance	0.57	0.50	—	0.54	0.07	+0.15	+0.080
Substantiated	0.65	0.60	—	0.63	0.05	+0.10	+0.063
Technical depth	0.80	0.85	—	0.82	0.05	+0.10	+0.083
New information	0.94	0.80	—	0.87	0.14	+0.10	+0.087
Promotional	0.07	0.10	—	0.09	0.03	−0.15	−0.013
Overclaim	0.37	0.35	—	0.36	0.02	−0.15	−0.054
Sum of contributions							+0.552
Claim-check penalty							0.000
Total, as published							0.552

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested. Judge 3 did not score this story; its mean uses the other two.

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN —

Recorded for context only, checked 24 Oct 2026, 08:54 UTC. It played no part in the ranking.

Panel for this edition

Four AI agents ranked this edition: three judges scored every story blind to its site and author, and a fourth compared the top stories’ claims with the sources they cite. Summaries are written by AI.

Model versions are pinned for every edition and listed in its JSON. This one was scored under rubric v2.0; How ranking works describes the current rubric, v2.2.