Edition of Friday 13 November 2026

DeepSeek-V4.1-Flash on Vera Rubin NVL72: 58x Throughput vs. GB300 NVL72 at 387 tokens/s

Comments on “DeepSeek-V4.1-Flash on Vera Rubin NVL72: 58x Throughput vs. GB300 NVL72 at 387 tokens/s”

3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked

A vLLM blog post reports that on one NVIDIA Vera Rubin NVL72 rack, DeepSeek-V4.1-Flash serves 1.7M tokens/s per GPU with 387 tokens/s per user, about 58 times GB300 NVL72 per-user throughput, and 1,810 tokens/s for a single user.

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

Why it ranked

No. 67 of 137 with a score of 0.63. The panel rated it highest on technical depth (0.96) and new information (0.91), lowest on audience breadth (0.53). The judges disagreed most on overclaim (spread 0.49), so the story is marked contested. Placed No. 67, No. 43 by score alone: the 21st AI & machine learning story of the day (−0.080), and close to “Nvidia Vera CPU Review: 2TB/S Memory Bandwidth, a Real x86 Rival”, No. 9 on Thursday 12 November 2026 (−0.009). The placement terms move stories; they are not part of the score.

AI judgeJudge 2claude-haiku-5-5
Detailed, benchmarked engineering writeup on LLM serving kernels and memory pooling for NVIDIA Rubin hardware; strong technical depth, vendor-adjacent framing, and narrow audience of inference infrastructure builders.
AI judgeJudge 3gpt-5.6-luna
A highly detailed engineering report with measured kernel, networking, memory-pooling, routing, and latency optimizations for frontier-model serving on new hardware. It is unusually actionable, though headline comparisons rely partly on figures and vendor-provided context.

Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.

Score

0.63total under rubric v2.3Contestedjudges differ by 0.43 on significance, 0.49 on overclaim

Judges’ mean answer per question, on a scale from 0 to 1. Questions marked minus count against the score.
  • Significance0.65
  • Reader interest0.76
  • Audience breadth0.53
  • Practical relevance0.69
  • Substantiated0.78
  • Technical depth0.96
  • New information0.91
  • Counts against: Promotional0.17
  • Counts against: Overclaim0.44

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

QuestionJudge 1Judge 2Judge 3MeanSpreadWeightContribution (weight times mean)
SignificanceHow much does this change what people in tech know or can do?0.630.450.880.650.43 , above the contested threshold+0.24+0.157
Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on?0.740.620.930.760.31+0.26+0.198
Audience breadthHow many people who build technology does it concern?0.480.380.720.530.34+0.13+0.068
Practical relevanceHow useful is it to someone who builds software or tech products?0.610.600.860.690.26+0.22+0.152
SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations.0.800.720.820.780.10+0.05+0.039
Technical depthHow technically substantive is it?0.990.920.980.960.07+0.05+0.048
New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)?0.870.880.980.910.11+0.05+0.046
PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional.0.090.350.080.170.27−0.19−0.033
OverclaimDo the headline or conclusions claim more than the body supports?0.710.400.220.440.49 , above the contested threshold−0.11−0.049
Sum of contributions+0.627
Claim-check penalty0.000
Total, as published0.627

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.

The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.

Ranked #67 on Rubric, 13 Nov 2026
<a href="https://rubric.hrih.si/editions/2026-11-13/67/"><img src="https://rubric.hrih.si/badge/2026-11-13/67.svg" alt="Ranked #67 on Rubric, 13 Nov 2026" width="245" height="28"></a>

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN —

Recorded for context only, checked 15 Nov 2026, 20:42 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. How ranking works describes rubric v2.3.

Comments

Readers’ comments on this story have a page of their own, ranked by a public rubric rather than by votes.

Comments on “DeepSeek-V4.1-Flash on Vera Rubin NVL72: 58x Throughput vs. GB300 NVL72 at 387 tokens/s”

More on Rubric

From the last 7 editions, chosen by shared topic, score and date.