ModelsNov 16, 2026 · 23:248 sources

Codex Coding Harness Ties for First Place on Ripple-65B Preview Benchmarks

From huggingnews.com

Image: huggingnews.com
AI;DR

The Codex coding harness tied for first place on the Hugging Face Harness Leaderboard when used with RicursiveAI's Ripple-65B Preview across 198 tasks. The Ripple-65B model is a dense large language model trained from scratch for $32M and features a 1M token context window.

Read the sourceSign in to suggest

Sources

+ 6 more
  1. SUPPORT@ricursiveai“Ripple was built on a $32 million compute budget, one-tenth the cost of the previous open-weight SOTA on agentic coding”x.com
  2. SUPPORT@teortaxestex“But it actually loses to V4 Flash 284B?”x.com
  3. SUPPORT@ricursiveai“our $32M R&D cost fully encompasses our pretraining ladder development, data mixing, evals, debugging, synthetic data generation, posttraining recipe development, and the final hero pre, mid, and posttraining run!”x.com
  4. SUPPORT@talaei_shayan“We trained it from scratch with a core team of five researchers, and it sets a new U.S. open-weight SOTA on Terminal-Bench 2.1 for agentic coding.”x.com
  5. SUPPORT@elliotarledge“I was not expecting a 65 billion parameter model to be performing as well as models over 10x its size on agentic coding tasks”x.com
  6. SUPPORT@eliebakouch“beautiful scaling ladder, interesting choices like value residual (interesting contribution pattern 👀) and.. SLIDING WINDOW ATTENTION AGAIN!”x.com
#codingMarkdown

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today