Writing Weights Index: Open-Source Models Struggle at Weight-Space Programming
eliebakouch · x · 2026-11-03
A Writing Weights Index was run from today's Weight-Space Programming post, and open-source models struggle on the task.
The index (100·q/(1+q), normalized against a PPO 500M baseline) shows GPT-6 Astra leading at 60.7, followed by Gemini 4 Carbon (57.7) and Claude Opus 5.5 (51.2). Open-weight models lag far behind: Qwen3.8 24T A9SB scores 12.5, DeepSeek V4.1 Flash 10.2 just 10.0, GLM-5.3 6.2, with MiniMax M3.1 Flash and Step 5 Preview near zero. Note: model names are as listed in the circulating chart; the post doesn't include full methodology.
More from Models
- Fabricated citations collapse from 55% to 0%: new models nearly fix the hallucinated-reference failure — RexDouglass · 2026-11-03
- Apple hints at monetizing third-party AI apps for new revenue on earnings call — TechCrunch AI · 2026-11-03
- Gemini 4 Argon ships, putting Google back in the frontier model race after a year-long lull — danshipper · 2026-11-03
- Hands-on audit punctures Astra's '0.016% weight programming' claim: it just disabled Qwen's thinking mode — teortaxesTex · 2026-11-03
- Leak: Unreleased astra6.2 builds a 3DS-style Mario Kart in 20 min under 1% quota — lxfater · 2026-11-03
- "Sonnet 5.5 used to be smarter": users sense capability regression — nlarusstone · 2026-11-03