WEBVTT

1
00:00:01.300 --> 00:00:07.721
This is Anthropic, Read Aloud. Today: Maximizing the value of your Claude Code sessions.

2
00:00:09.021 --> 00:00:11.801
Maximizing the value of your Claude Code sessions.

3
00:00:12.251 --> 00:00:17.841
Practical tips for how to run efficient sessions that get the most value from every token.

4
00:00:18.291 --> 00:00:22.344
From Claude Blog, published August 14, 2026.

5
00:00:23.344 --> 00:00:24.290
TL;DR.

6
00:00:24.740 --> 00:00:33.166
Run /clear between tasks. This prevents prior irrelevant context from being sent back to the model, which can reduce token usage.

7
00:00:33.406 --> 00:00:42.131
Set your model and effort level before you start. Changing either one mid-conversation can bust your prompt cache, which can increase token cost.

8
00:00:42.371 --> 00:00:51.697
@-mention files instead of naming them. The file gets attached to your message directly, which saves a Read call, or a search if Claude has to go find it.

9
00:00:51.937 --> 00:01:02.556
Add quiet flags to noisy commands, or run them in a subagent. Command output is added to the conversation just like a file, and stays there for the rest of the session.

10
00:01:02.796 --> 00:01:13.072
Run /context once in a fresh session. It shows what's loaded ( CLAUDE.md, M C P tool definitions), so you can cut out anything unnecessary.

11
00:01:13.312 --> 00:01:23.019
/compact before you take a break from your keyboard. The prompt cache expires after an hour, and summarizing a conversation is much cheaper while it's still cached.

12
00:01:23.819 --> 00:01:25.169
Maximizing value.

13
00:01:25.619 --> 00:01:38.735
Until pretty recently, the tools you wrote code with were a flat fee (or free). Your editor cost the same whether you fixed one test or fifty that afternoon, so an individual task didn't really have a price of its own.

14
00:01:39.185 --> 00:01:47.830
With agentic coding tools like Claude Code, it does. The same completed task can also cost different amounts depending on how you use it.

15
00:01:48.280 --> 00:02:08.045
In one session, Claude reads the test and the file it covers, makes the edit, and is done in a handful of turns. In another, it greps around the repo first, reads a dozen files on its way to the same two, and every one of those turns also drags along everything else that's been read into the conversation since this morning.

16
00:02:08.495 --> 00:02:16.971
It's the same fix, but you spent a different number of tokens on it, and the whole time the model was also having to think about ten files it didn't need.

17
00:02:17.421 --> 00:02:26.591
Being efficient with tokens doesn't mean using fewer of them overall. It means making sure the ones you do use go towards the thing you actually asked for.

18
00:02:27.041 --> 00:02:36.497
So let's look at what decides the price of a token, then what decides how many of them a session sends, and along the way, what that means for how you run a session.

19
00:02:37.297 --> 00:02:39.126
What decides the price of a token.

20
00:02:39.576 --> 00:02:51.683
You're billed per token, but what you're actually paying for is inference: the time it takes a G P U (or a T P U, or whatever the model happens to be running on) to run the model over your tokens.

21
00:02:52.133 --> 00:03:03.173
Three things decide how much of that time a token takes: which model you're running, whether it's an input token (going in) or an output token (coming out), and whether it was cached.

22
00:03:03.973 --> 00:03:04.600
Model.

23
00:03:05.050 --> 00:03:17.475
A bigger model does more work on both input and output tokens. Which model is worth it for which kind of work is a topic on its own, and we covered it in Choosing a Claude model and effort level in Claude Code.

24
00:03:17.925 --> 00:03:31.838
For this post, all you need to know is that everything else we're about to cover gets multiplied by the model's price: use a larger model when the problem is genuinely hard or ambiguous, and a smaller one when the work is routine.

25
00:03:32.288 --> 00:03:38.493
Figure 1: Curves are for illustration purposes only. They do not represent real benchmark data.

26
00:03:39.493 --> 00:03:56.406
That was Maximizing the value of your Claude Code sessions, from Claude Blog, published August 14, 2026. Read the original at claude dot com. This unofficial audio edition was generated automatically for the podcast Anthropic, Read Aloud.