[Linkpost] “Are the Costs of AI Agents Also Rising Exponentially?” by Toby_Ord

This is a link post.<p> There is an extremely important question about the near-future of AI that almost no-one is asking. </p><p> We’ve all seen the graphs from METR showing that the length of tasks AI agents can perform has been growing exponentially over the last 7 years. While GPT-2 could only do software engineering tasks that would take someone a few seconds, the latest models can (50% of the time) do tasks that would take a human a few hours.</p><p> As this trend shows no signs of stopping, people have naturally taken to extrapolating it out, to forecast when we might expect AI to be able to do tasks that take an engineer a full work-day; or week; or year.</p><p> But we are missing a key piece of information — the cost of performing this work. </p><p> Over those 7 years AI systems have grown exponentially. The size of the models (parameter count) has grown by 4,000x and the number of times they are run in each task (tokens generated) has grown by about 100,000x. AI researchers have also found massive efficiencies, but it is eminently plausible that the cost for the peak performance measured by METR has been [...]</p> <p>---</p><p><strong>Outline:</strong></p><p>(13:02) Conclusions</p><p>(14:05) Appendix</p><p>(14:08) METR has a similar graph on their page for GPT-5.1 codex. It includes more models and compares them by token counts rather than dollar costs:</p> <p>---</p> <p><b>First published:</b><br/> February 2nd, 2026 </p> <p><b>Source:</b><br/> <a href="https://forum.effectivealtruism.org/posts/AbHPpGTtAMyenWGX8/are-the-costs-of-ai-agents-also-rising-exponentially?utm_source=TYPE_III_AUDIO&utm_medium=Podcast&utm_content=Source+URL+in+episode+description&utm_campaign=ai_narration" rel="noopener noreferrer" target="_blank">https://forum.effectivealtruism.org/posts/AbHPpGTtAMyenWGX8/are-the-costs-of-ai-agents-also-rising-exponentially</a> </p> <p><strong>Linkpost URL:</strong><br><a href="https://forum.effectivealtruism.org/out?url=https%3A%2F%2Fwww.tobyord.com%2Fwriting%2Fhourly-costs-for-ai-agents" rel="noopener noreferrer" target="_blank">https://www.tobyord.com/writing/hourly-costs-for-ai-agents</a></p> <p>---</p> <p>Narrated by <a href="https://type3.audio/?utm_source=TYPE_III_AUDIO&utm_medium=Podcast&utm_content=Narrated+by+TYPE+III+AUDIO&utm_term=ea_forum&utm_campaign=ai_narration" rel="noopener noreferrer" target="_blank">TYPE III AUDIO</a>.</p> <p>---</p><div style="max-width: 100%";><p><strong>Images from the article:</strong></p><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/b5wbyy8b3ihkn7it3huq" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/b5wbyy8b3ihkn7it3huq" alt="Graph showing task duration versus LLM release date, titled "The time-horizon of software engineering tasks different LLMs can complete 50% of the time"." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/uysbgcowwob4th7zqz7n" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/uysbgcowwob4th7zqz7n" alt="Graph showing "Agent Performance on HCAST & RE-Bench by Cost (50% Time Horizon)" comparing AI models." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/oaejy9lxlz66cguxbq0w" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/oaejy9lxlz66cguxbq0w" alt="Graph showing agent performance on HCAST and RE-Bench by cost at 50% time horizon." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/jrb0ql43llkqdmnce5ge" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/jrb0ql43llkqdmnce5ge" alt="Graph showing agent performance on benchmarks by cost at 50% time horizon." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/kqzwckzwzqt414fcjseh" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/kqzwckzwzqt414fcjseh" alt="Graph showing "Agent Performance on HCAST & RE-Bench by Cost (50% Time Horizon)" comparing AI models." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/plmzvzejh7ao4roethyg" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/plmzvzejh7ao4roethyg" alt="Graph showing "Agent Performance on HCAST & RE-Bench by Cost (50% Time Horizon)" with AI models plotted." style="max-width: 100%;" /></a><hr style="margin-top: 24px; margin-bottom: 24px;" /><a href="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/zladjf4ape5irbq3hgee" target="_blank"><img src="https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/AbHPpGTtAMyenWGX8/zladjf4ape5irbq3hgee" alt="Line graph titled "Agent Performance on HCAST & RE-Bench by Token Count (50% Time Horizon)" showing multiple AI models' performance curves." style="max-width: 100%;" /></a><p><em>Apple Podcasts and Spotify do not show images in the episode description. Try <a href="https://pocketcasts.com/" target="_blank" rel="noreferrer">Pocket Casts</a>, or another podcast app.</em></p></div>