Claude Sonnet 5vsClaude Sonnet 4.5
Across 10 shared benchmarks, Claude Sonnet 5 leads overall: Claude Sonnet 5 wins 10, Claude Sonnet 4.5 wins 0, with 0 ties and an average score difference of +36.60.
Claude Sonnet 5
Anthropic · 2026-06-30 · Multimodal model
Claude Sonnet 4.5
Anthropic · 2025-09-30 · Chat model
Claude Sonnet 510 wins(100%)(0%)0 winsClaude Sonnet 4.5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 10 shared benchmarks.
Coding and Software Engineer
Claude Sonnet 5 3/3| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| Text Arena (Coding) | 1,54410 / 35Thinking High (No Tools) | 1,38732 / 35Normal (No Tools) | +157.16 |
| WeirdML v2 | 68.7820 / 52Thinking High (With Tools) | 46.7040 / 52Normal (With Tools) | +22.08 |
| SWE-bench Verified | 85.207 / 114极高强度思考（工具） | 828 / 114 | +3.20 |
Math and Reasoning
Claude Sonnet 5 2/2| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| FrontierMath v2 | 65.6117 / 34最高（无工具） | 23.8631 / 34Thinking (No Tools, 32K Budget) | +41.75 |
| FrontierMath Tier 4 v2 | 29.2716 / 34最高（无工具） | 2.4430 / 34Thinking (No Tools, 32K Budget) | +26.83 |
AI Agent - Information Search
Claude Sonnet 5 1/1| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| BrowseComp | 84.707 / 54Thinking (With Tools + Internet) | 24.1052 / 54 | +60.60 |
AI Agent - Tool Usage
Claude Sonnet 5 1/1| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| OSWorld-Verified | 81.206 / 26极高强度思考（工具） | 61.4022 / 26 | +19.80 |
Commonsense Reasoning
Claude Sonnet 5 1/1| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| SimpleBench | 57.9023 / 67Normal (No Tools) | 54.3027 / 67Normal (No Tools) | +3.60 |
General Evaluation
Claude Sonnet 5 1/1| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| GPQA Diamond | 90.5333 / 226极高强度思考（无工具） | 83.4099 / 226 | +7.13 |
General Knowledge
Claude Sonnet 5 1/1| Benchmark | Claude Sonnet 5 | Claude Sonnet 4.5 | Diff |
|---|---|---|---|
| HLE | 57.409 / 181极高强度思考（工具） | 33.6085 / 181 | +23.80 |
Specs
| Field | Claude Sonnet 5 | Claude Sonnet 4.5 |
|---|---|---|
| Publisher | Anthropic | Anthropic |
| Release date | 2026-06-30 | 2025-09-30 |
| Model type | Multimodal model | Chat model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 1M | 1000K |
| Max output | 128K | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Claude Sonnet 5 | Claude Sonnet 4.5 |
|---|---|---|
| Text input | $2 / 1M tokens | $3 / 1M tokens |
| Text output | $10 / 1M tokens | $15 / 1M tokens |
| Cache read | $0.3 / 1M tokens | $0.3 / 1M tokens |
| Cache write | $3.75 / 1M tokens | $3.75 / 1M tokens |
Summary
- Claude Sonnet 5leads in:Coding and Software Engineer (3/3), Math and Reasoning (2/2), AI Agent - Information Search (1/1), AI Agent - Tool Usage (1/1), Commonsense Reasoning (1/1), General Evaluation (1/1), General Knowledge (1/1)
On average across the 10 shared benchmarks, Claude Sonnet 5 scores 36.60 higher.
Largest single-benchmark gap: Text Arena (Coding) — Claude Sonnet 5 1,544 vs Claude Sonnet 4.5 1,387 (+157.16).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.