SVG generation test on Mistral Large 4 and other models
What happened
On October 7, 2026, replying on Hacker News to complaints that benchmarks are saturated, Simon Willison ran an SVG generation test with the llm tool while commenting on Mistral Large 4. He had claude-opus-5.5, gpt-6.1-sol, gemini-3.8-flash and mistral/mistral-large-4 each generate an SVG of an armadillo on Mars crossing the road in fishnet stockings, all at default reasoning levels. The post gives no output quality, comparison or scores for the models.
Written by AI from the coverage · updated 52 minutes ago
Coverage
Follow the reports to see the story from different sides.
- Simon WillisonMistral Large 4
Simon Willison 在 Hacker News 评论 Mistral Large 4 时，用 llm 工具让 claude-opus-5.5、gpt-6.1-sol、gemini-3.8-flash 和 mistral/mistral-large-4 分别生成"火星上穿网袜乱穿马路的犰狳"SVG，以回应"基准已饱和"的吐槽。各模型均使用默认推理等级。
Heat over time
Not enough continuous observations to draw a trend yet.