Fixing Tool-Call Accuracy in vLLM
Comments on “Fixing Tool-Call Accuracy in vLLM”
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 by Jesse Donkervliet and Junhao Li explains why tool-call schema accuracy matters for serving open-weight models in agentic workloads. They report that Kimi K2.6 on vLLM 0.28 produced schema-conforming tool calls only 92% of the time.
Why it ranked
No. 45 of 124 with a score of 0.63. The panel rated it highest on new information (0.90) and technical depth (0.81), lowest on significance (0.53).
Every answer behind this is in the table under Score.
Score
0.63total under rubric v2.2
- Significance0.53
- Reader interest0.73
- Audience breadth0.53
- Practical relevance0.80
- Substantiated0.74
- Technical depth0.81
- New information0.90
- Counts against: Promotional0.16
- Counts against: Overclaim0.26
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 | 0.52 | 0.45 | 0.62 | 0.53 | 0.17 | +0.24 | +0.127 |
| Reader interest | 0.60 | 0.72 | 0.88 | 0.73 | 0.28 | +0.26 | +0.191 |
| Audience breadth | 0.47 | 0.55 | 0.58 | 0.53 | 0.11 | +0.13 | +0.069 |
| Practical relevance | 0.83 | 0.75 | 0.82 | 0.80 | 0.08 | +0.22 | +0.176 |
| Substantiated | 0.65 | 0.72 | 0.84 | 0.74 | 0.19 | +0.05 | +0.037 |
| Technical depth | 0.86 | 0.80 | 0.78 | 0.81 | 0.08 | +0.05 | +0.041 |
| New information | 0.88 | 0.90 | 0.93 | 0.90 | 0.05 | +0.05 | +0.045 |
| Promotional | 0.11 | 0.20 | 0.18 | 0.16 | 0.09 | −0.19 | −0.031 |
| Overclaim | 0.28 | 0.25 | 0.25 | 0.26 | 0.03 | −0.11 | −0.029 |
| Sum of contributions | +0.626 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.626 | ||||||
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.
<a href="https://rubric.hrih.si/editions/2026-10-27/45/"><img src="https://rubric.hrih.si/badge/2026-10-27/45.svg" alt="Ranked #45 on Rubric, 27 Oct 2026" width="245" height="28"></a>Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 161 pts, front page
Recorded for context only, checked 27 Oct 2026, 05:46 UTC. It played no part in the ranking.