The AI writing landscape shifted again in 2024 as Anthropic and OpenAI released their most capable models to date. For teams evaluating or already using these tools, the updates bring meaningful improvements—and a few trade-offs to consider.
Claude 3.5 Sonnet: Long-Form and Structure
Anthropic's Claude 3.5 Sonnet, released in mid-2024, represents a significant step up in long-context handling and instruction-following. Early adopters report stronger performance on structured output—JSON, markdown, and multi-section documents—with better consistency across very long inputs (100K+ tokens).
For writers, the gains are most visible in PRDs, technical documentation, and legal summaries where structure and nuance matter. The model also shows improved "refusal" behavior, declining inappropriate requests more gracefully instead of producing off-brand or unsafe output.
GPT-4o and Omni: Speed and Multimodality
OpenAI's GPT-4o ("omni") brings native voice and vision to ChatGPT. For writing workflows, the headline improvements are faster inference—responses feel noticeably snappier—and improved handling of mixed inputs: text plus images, PDFs, or spreadsheets in a single conversation.
Teams using ChatGPT for research, document analysis, or design feedback will benefit from the multimodal capabilities. The model can now "see" screenshots, diagrams, and charts, making it useful for reviewing layouts, extracting data from tables, and comparing visual content.
Practical Takeaway
Both releases push toward longer context windows and better instruction adherence. If you're evaluating tools, don't rely on benchmark scores alone. Test with your real documents, your real prompts, and your real quality bar. The best model for your use case may not be the one with the highest MMLU score.
Discussion
Sign in to comment. Your account must be at least 1 day old.