Claude 3.5 Sonnet
Mid-Tier Surpasses Prior Flagships
A mid-tier Claude 3.5 model that, on release, beat the previous flagship (Claude 3 Opus) on most benchmarks while charging 1/5 the price. Cemented Anthropic's reputation as the leading model for code generation — the dominant Claude model in IDEs like Cursor and Windsurf throughout 2024-25.
Intelligence
Below avg
Cost
Moderate
$1.00 in / $4.00 out
Context
200K
Up to 200,000 tokens
How are Intelligence, Speed & Cost bucketed?
Intelligence and Speed buckets are percentile ranks on
Artificial Analysis. Cost buckets are fixed dollar
thresholds keyed off output-token price ($/M out).
Intelligence
- Top 1%≤ 1%
- Top 5%≤ 5%
- Top 10%≤ 10%
- Good≤ 25%
- Medium≤ 50%
- Below avg> 50%
Speed
- Top 1%≥ 345 tok/s
- Top 5%≥ 237 tok/s
- Top 10%≥ 196 tok/s
- Good≥ 146 tok/s
- Medium≥ 90 tok/s
- Slow< 90 tok/s
Cost
- Freeopen weights · self-host
- Low< $1 / M out
- Moderate$1–5 / M out
- High≥ $5 / M out
Why it matters
Claude 3.5 Sonnet established that "mid-tier > prior flagship" at cheaper pricing was the new release cadence — meaning customers should expect 6-month obsolescence of the model they're building on. This reframed enterprise procurement: build on the model abstraction (provider, not version) and assume continuous quality upgrade.
Core Capabilities
Long Documents
Handles entire codebases, books, and multi-doc RAG.
Multimodal
Combines text, vision, and audio in one model.
Generative
Produces images, video, audio, or other media.
Agent Workflows
Built for tool use and autonomous tasks.
Context Window
200k tokens
≈ 154+ pages
4k Chat 聊天
32k Long docs 长文档
128k Books 整本书
400k Multi-doc 多文档
1M Codebase 整个代码库
10M
200k
Availability
API
Available
Product / App
Available
Open Source
Not released
Enterprise
Contact sales
Pricing Model
Pay per token
Input and output billed separately.
Pay-per-token Capability / Performance
Where this model sits relative to the middle 60% of models in the tree. All scores are 0–10 (higher is better).
Lower 20% Upper 80% This model
Context / memory
Context window size · log-scaled
6.0
9.0
6.7
Lower 20% 20th percentile — 20% of models score below this This model Where the current model lands Upper 80% 80th percentile — only 20% of models score above this
Percentile boundaries are computed across every model in the tree that reports the underlying benchmark for each capability.
What it feels like
- The model that made AI coding actually trustworthy — produces working code nearly every time
- 49% on SWE-bench Verified at release; ahead of o1-preview and significantly past prior versions (33%)
- Outperformed the larger Claude 3 Opus on most benchmarks while costing 80% less
- 92.0% on HumanEval — edges GPT-4o (90.2%) on Python function tests
- Multi-step task completion 40-54% without human help — over half of long tasks still need course-correction
- Some users report needing more back-and-forth on long-form outputs
Best use cases
- Front-end and back-end web development — the canonical 2024 coding copilot
- Refactors and feature work where reliability beats raw IQ
- Tools-and-agents stacks (Cline / Claude Dev / Cursor) — became the default model
- GitLab DevSecOps pipelines (10% reasoning gain, no slowdown)
Tools to try
Not ideal for
- Languages or frameworks outside its training distribution
- Multi-hour autonomous tasks — prefer Claude 4 family
- Hardest reasoning / math benchmarks — better on o1 / o3
Model Evolution
Claude 3
Mar 2024
Claude 3.5 Sonnet
Jun 2024
Model Context Protocol
Nov 2024
Claude 3.7 Sonnet
Feb 2025