Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

ReViSQL-K2.6

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

0

Findings

1

Edges

2

Sources

1

Corpus findings

2026-08-30 / thought-leaders-researcherThinking Machines Lab and UIUC beat the human text-to-SQL benchmark for the first time, with no pipeline engineeringAn August 27 post from Thinking Machines Lab with Yuxuan Zhu, Tengjun Jin, Yoojin Choi and Daniel Kang describes ReViSQL-K2.6, an RLVR run on Kimi-K2.6 using VeriEQL semantic-equivalence verification plus rule-based process rewards on expert-curated data. With 16-sample self-consistency it reaches 92.97% on Arcwise-Plat-SQL at $0.56/task, edging the 92.96% human benchmark, and greedy decoding gets 91.37% at $0.035/task. The load-bearing claim for builders is that it outperforms scaffolded baselines by 8 to 22 points while deleting the orchestration layer entirely.

Graph relationships

USES
Kimi K2.6
ReViSQL-K2.6 is an RLVR run on Kimi-K2.6 using semantic-equivalence verification.
Source finding
BENCHMARKED AGAINST
Kimi K2.6
ReViSQL-K2.6 achieved 92.97% on the Arcwise-Plat-SQL benchmark, edging the human benchmark of 92.96%.
Source finding

Source trail

Thinking Machines Lab

Graph sources

entity graphfindings textkg edgeskg entities