Ask Me Anything

JonathonAiThe Ai Models I Use For Work

The AI models I Use for Work

Published Sep 4, 2025
2 minutes read

Here are some rough notes on the AI models I use as of August 26th, 2025.

Important Caveats

Search & Answer Engines

ToolUse Case
KagiMy primary search engine. Think "Google, but significantly better" with no ads and customizable results.
Claude Opus 4.1Excellent as an "answer engine" that can search the web and provide quick but thorough answers to "stupid" and complex questions.
Gemini Advanced Deep ResearchSeems to produce reports I appreciate more than OpenAI's.
Claude ResearchI'm currently exploring this and like the breadth of searches coupled with shorter final reports than other providers. Verdict still TBD.

Writing & Content Creation

ToolUse Case
Claude Opus 4.1My go-to for general copywriting help, creative writing tasks, translation, and editing existing text. It is the only model that tells me when my writing and ideas are bad. I find this refreshing.
Gemini 2.5 ProMy go to for creating structured documents like guides, proposals, and grant applications.

Image Generation

ToolUse Case
GPT-4o & Nano Banana (Gemini)Useful for quick edits, brainstorming visual ideas, and initial explorations.
MidjourneyMy preferred choice for any serious image generation when final quality, artistic control and taste matter.

"AI Chores" (Utility Tasks)

ToolUse Case
GPT-5-miniGreat for quick tasks like reformatting text, reorganizing information and other minimally complex data manipulation.
Gemini 2.5 ProMy choice for summarizing lengthy texts, synthesizing research from multiple documents, and structuring longer documents.

Powering AI Helpers

A Combination Approach: I often use a mix of OpenAI o3 Pro, Gemini 2.5 Pro, Claude Opus 4.1 and Grok 4 to power different aspects of my AI advisors or personas, depending on the specific task the persona is undertaking (e.g., research, writing, critical analysis). Each model has its own quirks, but using them together helps surface more insights and perspectives. Multi-modelling has been as helpful as it has been time consuming. Sometimes I feel like I'm the bottleneck in the process.

Coding

ToolUse Case
v0For generating quick UI prototypes or landing page mockups using text prompts and scribbled wireframes.
CursorFor any serious development work, I use Cursor for its code autocomplete.
Claude CodeFor more complex tasks, Claude code with Opus 4.1 has completely replaced my use of Cursor's coding agent. Usage limits are much higher and results seem to be better so far. Sonnet 4 is excellent for implementation, especially for smaller, well-scoped, but annoying tasks.

Feel free to use this list as a starting point, but please experiment and discover what works best for you!

PS: The next cohort of my AI literacy for business course starts on September 16th in case you'd like structured guidance on how to implement these tools to support your work.