Claude Opus 4.8vsOpus 4.5
Across 10 shared benchmarks, Claude Opus 4.8 leads overall: Claude Opus 4.8 wins 10, Opus 4.5 wins 0, with 0 ties and an average score difference of +21.67.
Claude Opus 4.8
Anthropic · 2026-05-28 · Reasoning model
Opus 4.5
Anthropic · 2025-11-25 · Reasoning model
Claude Opus 4.810 wins(100%)(0%)0 winsOpus 4.5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 10 shared benchmarks.
Coding and Software Engineer
Claude Opus 4.8 3/3| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| Text Arena (Coding) | 1,5459 / 35Normal (No Tools) | 1,47920 / 35Normal (No Tools) | +66.05 |
| SWE-bench Verified | 88.605 / 114Extended (with tools) | 80.909 / 114Extended (with tools) | +7.70 |
| WeirdML v2 | 70.4518 / 52Normal (With Tools) | 63.7024 / 52Thinking (With Tools, 16K Budget) | +6.75 |
General Knowledge
Claude Opus 4.8 2/2| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| HLE | 57.908 / 181Extended (with tools) | 43.2052 / 181Extended (with tools) | +14.70 |
| LiveBench | 78.794 / 115Deep Thinking (No Tools) | 75.9611 / 115Thinking (No Tools, 64K Budget) | +2.83 |
Math and Reasoning
Claude Opus 4.8 2/2| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| FrontierMath Tier 4 v2 | 56.109 / 34最高（无工具） | 4.8828 / 34Thinking (No Tools, 32K Budget) | +51.22 |
| FrontierMath v2 | 809 / 34最高（无工具） | 34.3930 / 34Thinking (No Tools, 32K Budget) | +45.61 |
AI Agent - Tool Usage
Claude Opus 4.8 1/1| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| MCP-Atlas | 82.206 / 38Deep Thinking (With Tools) | 69.8026 / 38Thinking High (With Tools) | +12.40 |
Commonsense Reasoning
Claude Opus 4.8 1/1| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| SimpleBench | 64.8010 / 67Normal (No Tools) | 6214 / 67Extended (no tools) | +2.80 |
General Evaluation
Claude Opus 4.8 1/1| Benchmark | Claude Opus 4.8 | Opus 4.5 | Diff |
|---|---|---|---|
| GPQA Diamond | 93.6011 / 226Thinking High (No Tools) | 8771 / 226Extended (no tools) | +6.60 |
Specs
| Field | Claude Opus 4.8 | Opus 4.5 |
|---|---|---|
| Publisher | Anthropic | Anthropic |
| Release date | 2026-05-28 | 2025-11-25 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 1M | 200K |
| Max output | 125K | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Claude Opus 4.8 | Opus 4.5 |
|---|---|---|
| Text input | $5 / 1M tokens | $5 / 1M tokens |
| Text output | $25 / 1M tokens | $25 / 1M tokens |
| Cache read | $0.5 / 1M tokens | $0.5 / 1M tokens |
| Cache write | $6.25 / 1M tokens | $6.25 / 1M tokens |
Summary
- Claude Opus 4.8leads in:Coding and Software Engineer (3/3), General Knowledge (2/2), Math and Reasoning (2/2), AI Agent - Tool Usage (1/1), Commonsense Reasoning (1/1), General Evaluation (1/1)
On average across the 10 shared benchmarks, Claude Opus 4.8 scores 21.67 higher.
Largest single-benchmark gap: Text Arena (Coding) — Claude Opus 4.8 1,545 vs Opus 4.5 1,479 (+66.05).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.