Anthropic cost guide: tune effort first, then weigh model tiers
What happened
On October 12, Computing Life published a read of Anthropic's cost guide. The guide says to compare models by cost per successful task, test effort adjustments first, and only then judge whether splitting work across model tiers pays off. Switching models mid-session can break the prompt cache, while changing effort does not. On SWE-bench Pro, running low first and rerunning failures at high scored 97.0% at $0.17 per task; using high throughout scored 95.3% at $0.29 per task.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- Computing Life · Share · YagePickAnthropic cost guide: tune effort first, or switch models?
Anthropic's cost guide says to compare models by cost per successful task, and to test effort settings before deciding whether a multi-model tier is cheaper. Switching models mid-conversation can break prompt caching, while changing effort does not. On SWE-bench Pro, running low effort first and rerunning failures at high effort scored 97.0% at $0.17 per task; running high effort throughout scored 95.3% at $0.29.
Heat over time
Heat now 8·Comparable peak 9（Oct 13 12:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.