Skip to content
Tagged articles

software engineering benchmarks

2 articles · Page 1 of 1
ITPUB
ITPUB
Oct 10, 2026 · Artificial Intelligence

Google Tests Gemini 4 'Carbon' Internally as Argon Launch Stalls: 'Feels Like Opus 5.5'

Google is internally testing a new Gemini 4 checkpoint codenamed Carbon, which employees say matches Anthropic's Opus 5.5 for coding, while the already-announced Argon model remains unavailable to most users despite strong benchmarks in long-context software engineering but weaknesses in terminal interaction.

AI agentsArgonCarbon
0 likes · 11 min read
Google Tests Gemini 4 'Carbon' Internally as Argon Launch Stalls: 'Feels Like Opus 5.5'
Machine Heart
Machine Heart
Aug 2, 2026 · Artificial Intelligence

VeriLoop Coder‑E1 Open‑Source: Evidence‑Governed Spiral Drives Verified Recursive Self‑Improvement

VeriLoop Coder‑E1, an open‑source code model built on Qwen3.6‑27B, achieves top rankings on multiple software‑engineering benchmarks, leverages narrow‑domain PEFT fine‑tuning and a Self‑Harness execution loop, and introduces an evidence‑governed spiral that enables verified recursive self‑improvement while the community rapidly creates low‑precision quantizations.

Coder-E1Evidence-Governed SpiralPEFT
0 likes · 17 min read
VeriLoop Coder‑E1 Open‑Source: Evidence‑Governed Spiral Drives Verified Recursive Self‑Improvement