Retrospective: Nous' First Model gpt4-x-vicuna-13b Trained on 180k GPT-4 Outputs
Teknium · x · 2026-08-27
Nous Research released its first model, gpt4-x-vicuna-13b. This 13B model is based on Vicuna and was trained for 5 epochs on over 180k censor-free GPT-4 outputs. The datasets were contributed by Nous Research, Teknium, WizardLM, and GPT-4-LLM, with compute sponsored by Redmond AI. The team noted that the next step was to clean Vicuna to fully remove OpenAI censorship.
More from Models
- GLM 5.3 Flash Benchmark: Hits 881 tok/s on Dual DGX — teortaxesTex · 2026-08-27
- Rumor: Her-like ChatGPT Voice Mode with GPT-6 Coming This Year — flowersslop · 2026-08-27
- GLM 5.3 Flash offers 90% discount, pricing drops to $0.04 per 1M output tokens — shensi · 2026-08-27
- Developers discuss LLM coding degradation at long context — jazz788 · 2026-08-27
- Dry 5.6 Sol is sycophantic in chat but not when goal-oriented — Sauers_ · 2026-08-27
- Qwen Pruning Tests: 256 Experts Optimal, Q4 Beats Q2 — EyalToledano · 2026-08-27