Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

Frontis-MA1

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

Briefing refs

0

Findings

1

Edges

5

Sources

1

Corpus findings

2026-07-31 / arxiv-researcherFrontis-MA1 (35B) Reaches 71.21% MLE-Bench Lite Medal Average on a Single 12 GB RTX 4090, Weights ReleasedFrontis-MA1 (arXiv 2607.28568, July 30) is a 35B meta-evolution agent post-trained on OpenMLE, an open full-stack system for recursive self-improvement research spanning verifiable task environments (OpenMLE-Gym), operator learning (OpenMLE-RL), and long-horizon search (OpenMLE-Evo). Four atomic program-evolution operators — Draft, Improve, Debug, Crossover — are trained via execution-grounded SFT and RL, then composed into search. Under a 12-hour per-task budget on one RTX 4090 capped at 12 GB VRAM it lifts MLE-Bench Lite Medal Average from 39.39% to 60.61%, and to 71.21% with OpenMLE-Evo-Max, exceeding GPT-5.5 + Codex and approaching GPT-5.6 Sol and the 2.8T Kimi K3; model weights and the full stack are public.

Graph relationships

USES
OpenMLE
Frontis-MA1 is trained on OpenMLE open full-stack system
Source finding
BENCHMARKED AGAINST
Kimi K3
Frontis-MA1 approaches Kimi K3 on MLE-Bench performance
Source finding
BENCHMARKED AGAINST
GPT-5.6 Sol
Frontis-MA1 approaches GPT-5.6 Sol on MLE-Bench Lite Medal Average
Source finding
BENCHMARKED AGAINST
Codex
Frontis-MA1 exceeds GPT-5.5 plus Codex on MLE-Bench
Source finding
BENCHMARKED AGAINST
GPT-5.5
Frontis-MA1 exceeds GPT-5.5 on code evolution tasks
Source finding

Source trail

Graph sources

entity graphfindings textkg edgeskg entities