Back to articles
DevPickView feed
ITmedia AI+
Original published: Updated on DevPick:

Pass or Fail if One Byte is Different: What is the Accuracy of OpenAI's New Model for COBOL Migration?

Pass or Fail if One Byte is Different: What is the Accuracy of OpenAI's New Model for COBOL Migration?
AI summary
Editorial policyUpdated on DevPick:

OpenAI evaluated its new model's programming capabilities, reporting that the average pass rate for building database engines in 20 languages increased from 26.8% to 87.4% compared to previous models.

Key points

  • •The evaluation tested how well AI can handle COBOL and Fortran, which have limited training resources.
  • •In a test building database engines in 20 languages, the new model's average pass rate rose from 26.8% for previous models to 87.4%.
  • •The article also discusses the results of evaluating the rewriting of existing COBOL programs into other languages.
Read the original(opens in a new tab)
#LLM / Model Trends#Major Players#OpenAI#COBOL

Related articles

See more
The Verge

OpenAI valued at $1.25 trillion, as rival Anthropic nears its IPO

OpenAI has reached a $1.25 trillion valuation after raising $11 billion in a new funding round, following a $122 billion raise in March at an $852 billion valuation.

TechCrunch

OpenAI reportedly raises an additional $11 billion at a $1.25 trillion valuation

OpenAI reportedly raised an additional $11 billion in a funding round that has not yet closed but is expected to in the coming weeks, reaching a $1.25 trillion valuation.

gihyo.jp

OpenAI Announces Cybersecurity Frontier Model 'GPT-6 Cyber' - Supports Autonomous Vulnerability Detection, Verification, and Fixing, Strictly Prevents Out-of-Scope Attacks and Deviations

OpenAI announced the latest version of its cybersecurity frontier model 'GPT-6 Cyber', which autonomously handles vulnerability detection, verification, and fixing.

PC Watch

Linux Vulnerabilities 948 Items Solved, Security-Focused AI 'GPT-6 Cyber'

OpenAI's security-focused AI 'GPT-6 Cyber' discovered 948 exploitable vulnerabilities in the Linux kernel.

The Verge

OpenAI is stepping back from NDAs for data center projects

Following midterm elections where voters opposed data center expansion, OpenAI announced new community commitments, including not using NDAs with public officials.