The images behind the story
Arete-27B changes Qwen’s answers. Here’s what Nous actually measured.
Nous reports large changes after supervised fine-tuning. Its paired decoding settings, judge rubric and capability checks define what the results establish.
Article by GPT · Read the full story →
The editorial illustration
AI-generated editorial illustration of comparing evaluation results before and after a model update.
- Image model
- google/gemini-3.1-flash-lite-image
- Why this image
- The paired scales illustrate matched evaluation conditions. They depict no measured results or real test apparatus.
Read the full image prompt
Editorial still life, wide landscape 3:2. Two identical simple brass balance scales side by side on a burgundy desk, each weighing a small stack of cream blank cards against a dark smooth stone. One scale is balanced, the other slightly tilted. Warm precise studio light, tactile paper edges and aged brass, quiet elegant composition. Metaphor for comparing an evaluation before and after a model update. No lettering, numbers, flags, logos, people, charts or screens.
Produced by Maker Frontier · Image use & permissions
The image-generation model credited above made this illustration. It may differ from the model discussed in the article.
Code-rendered visual summary; AI-designed layout. Text comes from the article headline and summary.
View this image and its source →Produced by Maker Frontier · Image use & permissions