New AI models may be starting to sound too similar
A firsthand comparison of recent AI models found that both API models and can start to answer in increasingly similar ways. The pattern appears more clearly after several turns of conversation or when the task moves into a narrow subject area.
The outputs do not fully break down, but their rhythm, cautious wording, and repeated weak spots begin to look alike. One possible cause is that different models may now share the influence of similar .
The proposed working name is EchoCreep. Useful next checks include finding evaluation metrics for this sameness, testing whether on human-curated data reduces it, and seeing whether the pattern grows across newer model .
Key points
- Recent API models and may produce answers that feel increasingly alike.
- The sameness appears more after longer conversations or niche tasks.
- A possible cause is shared influence from .
- Human-curated is suggested as something worth testing.
- Agent builders should include longer multi-turn checks when comparing model quality and cost.