Context Budget

---
name: context-budget
description: Estimate the token cost of reading a file or directory before doing it. Use when asking "how many tokens would this cost", "how big is this file in tokens", "is this file too big to read", before reading large files, or when weighing whether to read something with limited context budget remaining.
license: MIT
metadata:
  author: AfterRealm
  version: "1.9.1"
---

# Context Budget — Token Estimation

Estimate how many tokens a file or directory will consume before reading it.

## How to Estimate

**Rule of thumb:** 1 token ≈ 4 characters (English text/code). This is a rough estimate.

**Opus 4.7 tokenizer note:** Opus 4.7 uses a new tokenizer that can produce up to **35% more tokens** for the same text vs. prior models. If the user is on Opus 4.7, multiply the rough estimate by ~1.35 for a conservative upper bound. Opus 4.6, Sonnet 4.6, and Haiku 4.5 use the older tokenizer.

When the user asks about a file or directory:

1. Get the file size in bytes using:
```bash
wc -c < "filepath"
```
Or for a directory:
```bash
find "dirpath" -type f -exec wc -c {} + | tail -1
```

2. Estimate tokens: `file_size_bytes / 4` (add ~35% headroom if they're on Opus 4.7)

3. Show the estimate with context AND model comparison:

```
File: path/to/file.py
Size: 12,400 bytes
Est. tokens: ~3,100 (~4,200 on Opus 4.7)
Context impact: 0.3% of 1M window

Cost to read (first time / cached):
  Opus 4.7: $0.016 / $0.002
  Opus 4.6: $0.016 / $0.002
  Sonnet 4.6: $0.009 / $0.001
  Haiku 4.5: $0.003 / $0.0003
```

Use these prices per 1M tokens for the comparison:

| Model | Input | Cache Read | Cache Write (5m) | Output |
|-------|-------|------------|------------------|--------|
| Opus 4.7 | $5.00 | $0.50 | $6.25 | $25.00 |
| Opus 4.6 | $5.00 | $0.50 | $6.25 | $25.00 |
| Sonnet 4.6 | $3.00 | $0.30 | $3.75 | $15.00 |
| Haiku 4.5 | $1.00 | $0.10 | $1.25 | $5.00 |

Opus 4.7 and 4.6 share the same per-token pricing. The 1M-context variants (`claude-opus-4-7[1m]`, `claude-opus-4-6[1m]`, `claude-sonnet-4-6[1m]`) are also at standard pricing — no long-context premium.

- "First time" = input price (first read, not yet cached)
- "Cached" = cache read price (subsequent reads in same session)
- Always show all four models so users can compare

For directories, also show file count and breakdown of largest files.

## Context Impact Table

| Tokens | % of 1M | % of 200K | Verdict |
|--------|---------|-----------|---------|
| < 1K | 0.1% | 0.5% | Trivial |
| 1-5K | 0.1-0.5% | 0.5-2.5% | Light |
| 5-20K | 0.5-2% | 2.5-10% | Moderate |
| 20-50K | 2-5% | 10-25% | Heavy — consider reading specific sections |
| 50K+ | 5%+ | 25%+ | Very heavy — read only what you need |

## Tips

- Suggest `--offset` and `--limit` on Read for large files
- For directories, suggest reading only the files that matter
- Minified JS/CSS files are token-dense — warn about them
- Binary files waste tokens — flag them