PulseAugur
EN
LIVE 11:48:24

Open-source PRAXIST AI system outperforms Claude Code on MLE-bench for 12x less cost

By PulseAugur Editorial · [1 sources] ·

An open-source AI research system named PRAXIST has achieved superior results on the MLE-bench compared to a Claude Code baseline. PRAXIST secured 49 gold medals for approximately $3,000 in compute costs, while the Claude Code baseline achieved 34 gold medals at a cost of $38,000. The system's success is attributed to its method of inheriting evidence from failed experiments rather than just scores, allowing lessons to be passed to subsequent generations. AI

IMPACT Demonstrates a novel approach to AI research that significantly reduces compute costs while improving performance, potentially influencing future AI development methodologies.

RANK_REASON The item describes a research system's performance on a benchmark, detailing its methodology and results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — mastodon.social →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Open-source PRAXIST AI system outperforms Claude Code on MLE-bench for 12x less cost

How we ranked this

Signal score
15 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The item describes a research system's performance on a benchmark, detailing its methodology and results. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Mastodon — mastodon.social TIER_1 English(EN) · gitgem ·

    An open-source AI research system just won 49 golds on MLE-bench for ~$3,000 in compute. The Claude Code baseline: 34 golds for $38,000. Same tasks, same rules,

    An open-source AI research system just won 49 golds on MLE-bench for ~$3,000 in compute. The Claude Code baseline: 34 golds for $38,000. Same tasks, same rules, 12x cheaper. The secret? It inherits evidence, not scores. Failed experiments pass their lessons to the next generation…

RELATED ENTITIES

RELATED TOPICS