Google's Gemini 3.7 Flash AI Model Shines in Coding, Falls Short in Creative Writing

Google recently released Gemini 3.7 Flash, its latest AI model, which boasts impressive capabilities for everyday tasks. Available in over 160 countries, this affordable model can process up to a million input tokens and generate results in various formats, including text, images, video, audio, and PDFs.

When compared to previous models, such as Gemini 3.6 Flash, the new version shows significant improvements. In coding tasks, it can produce working software with minimal errors, unlike its predecessor that required human intervention to fix its output.

In creative writing, Gemini 3.7 Flash generated a decent story, but still fell short of expectations. It struggled to maintain a paradox and write a sentence without revealing the plot machinery. In contrast, Qwopus3.5-27B-v3, a community fine-tuned model, produced a more cohesive narrative.

In terms of associative thinking, Gemini's performance was also mixed. While it generated some impressive imagery, it failed to dissolve the argument into a description of a lettuce seamlessly. GPT-5.6 Sol outperformed Gemini in this task by producing a more integrated and logical explanation.