Anthropic · The Long-Context Stretch
Claude 3.5 Sonnet
The model that made Claude the coding default. Coding benchmarks beat GPT-4o at launch.
By C.W. Jameson · Published 19 May 2026 · Last reviewed 19 May 2026
Claude 3.5 Sonnet was released in June 2024 and immediately posted the strongest SWE-bench verified numbers of any model. Operators who had kept Claude as a backup quietly made it primary. The version also introduced the first public computer-use capability in October 2024.
Field signature
Writes code with unusually complete error handling. Rarely leaves TODO stubs.
Specifications
| Released | 2024-06 |
|---|---|
| Context window | 200,000 tokens |
| Pricing | $3 / $15 per million tokens (historical; superseded by 4.x) |
| Modalities | text · image · PDF |
| License | Commercial API only |
| Era | The Long-Context Stretch |
Strengths
- Coding benchmarks
- First computer-use integration
- Refusal quality
Weaknesses
- Superseded by 4.x generation
Authentication markers
The fingerprints by which Claude 3.5 Sonnet can be identified from its output alone.
| Tell | Meaning |
|---|---|
| Will mention 'artifacts' in Claude.ai context. | Claude 3.5 era. |
Notable works
- SWE-bench verified leader June 2024
- First Claude computer-use
Market position
$3-$15 per million tokens
Partner offer
Anthropic's Claude family is the model lineage most operators end up on for serious agent work. The free tier remains useful.
Try Claude →Affiliate link — see disclosure.
Primary sources
From the Almanac shop
The Operator's Compendium
Every agent harness, every routing pattern, every cost trick. 90-page PDF.
$29 — Coming soon