What is the best AI for your resume and cover letter? This is the second run of my resume scoring model bake-off. I did two things this time. First, a catch-up run of 18 models that are still widely used, including versions a step or two back that are often cheaper to use on a free plan or a $20 a month plan, plus more open weight models. Second, a head-to-head of brand new, state-of-the-art, expensive models that dropped in the last week or so. Every model got the same prompt and the same inputs for our test candidate, Alex Clearwater, a PhD and postdoc moving into industry pharmacology, applying to a fake job at a fake company. I scored each resume and cover letter on the headline, highlights, formatting and organization, tailoring, and the cover letter. Then two AI judges (Claude Opus 5.5 and GPT-5.6 Sol) scored accuracy, checking every bullet against the template resume for anything made up or overstated, and Jev checked 13 yes or no rules. The AI models tested (30 in total across both parts): Catch-up run, 18 older models: • OpenAI: GPT-5.6 Sol, GPT-5.5, GPT-5.4 • Anthropic (Claude): Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Sonnet 4.6 • Google: Gemini 3.7 Flash, Gemini 3.5 Flash, Gemini 3.1 Pro Preview • xAI: Grok 4.6, Grok 4.5 • Open models: DeepSeek V4 Flash, DeepSeek V3.2, Llama 4 Maverick (Meta), Qwen 3.8 Max Prime, Kimi K3 (Moonshot), GLM 5.3 (Z.ai) New run, 6 new models: • OpenAI: GPT-6.1 Sol • Google: Gemini 4 Argon • Meta: Muse Wonder • Mistral: Mistral Large 4 • Qwen (Alibaba): Qwen 3.8 Max • DeepSeek: DeepSeek V4.1 Flash From Part 1, shown for comparison: • GPT-6 Sol, Claude Opus 5.5, Claude Sonnet 5.5, Gemini 3.8 Flash, Grok 4.7, DeepSeek V4 Pro In this episode you'll learn: • How older, cheaper, and open weight models compare to the newest AI models for resumes and cover letters • Why a resume can look great and still have major accuracy problems • Why some models write well but aren't accurate, and vice versa • When it might be worth using a different model for your cover letter • The cover letter habits I flag: telling the company what it wants, language that sounds like a weakness, passive closers, and very long sentences • Whether newer models are better at following instructions See every output side by side, plus the job, the template resume, and the example tailored resume and cover letter, at models.biotechcareercoach.com. It also has a guide to which model to use if you're a Claude user, a ChatGPT user, or want an open weight model. https://models.biotechcareercoach.com/ Missed Part 1? Watch it first to see how I set up the whole model bake-off test: Best AI for Resume and Cover Letter, Which Model Wins? [LINK] Want my help with your resume and cover letter? Join our Biotech Career Coach Skool community and work with me directly. It is $9 a month with a 7-day free trial. Learn more about the Collaboratory Career Hub community and access our free resources: Join our Skool Community Take the Free 7-day Interview Sprint Challenge Check out our sister podcast: Building Biotechs Send Carina a connection request on LinkedIn! Stay connected with us: 🌐 Website 📱 Facebook 📷 Instagram 💼 LinkedIn Chapters 00:00 Best AI for resume and cover letter, Part 2 01:03 How the AI resume test works 01:39 How I score the resume headline, highlights, and tailoring 02:56 Catch-up run, 18 older and open weight AI models 09:30 AI cover letter mistakes that sound arrogant or weak 11:27 What a strong AI cover letter opener looks like 16:05 Why long sentences hurt your cover letter 18:03 The reveal for the catch-up run 18:43 AI judges check resume accuracy and hallucinations 19:49 Jev rule checks for instruction following 22:16 Why a great looking AI resume can still be inaccurate 24:15 Older AI models vs last week's frontier models 25:29 Newest AI models head to head 30:23 The reveal for the newest models 32:15 Human scores vs AI judges on the new models 33:28 Compare AI resume outputs side by side 34:46 Should you combine AI models for resume and cover letter
