See key specs and per-benchmark scores for each model/mode. Scroll horizontally for all columns. 当前对比 2 个模型的评测数据与核心参数。
Claude Opus 4.6
Anthropic
Claude Opus 4.8
Anthropic
Each axis is a category average, normalized to a 100-point radar.
Relative edge: 常识推理 +2.8 / Relative gap: 数学推理 -21.7
Relative edge: 数学推理 +21.7 / Relative gap: 常识推理 -2.8
Method: for each model and benchmark, the chart first averages all scores in the current mode scope instead of taking the best score, then averages those benchmark scores within each category. Only benchmarks with at least two selected models scored are included; missing values are not counted as zero.
Best overall
Claude Opus 4.8 · 329.82
Best single
Claude Opus 4.8 · GDPval-AA 1890.00
Modality coverage
Claude Opus 4.6 · 2 modalities
Head to head
13
Benchmarks
2
Wins
11
Losses
-27.15
Average diff
Compare benchmark results across thinking modes and tool usage.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
Complete scores for each model/mode across selected benchmarks.
13 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Claude Opus 4.6 | Claude Opus 4.8 |
|---|---|---|
HLE 综合评估 | 53.00Extended Thinking ｜ Tools | 57.90Extended Thinking ｜ Tools |
LiveBench 综合评估 | 76.33Thinking Level · High | 78.79Deep Thinking Mode |
SWE-bench Verified 编程与软件工程 | 80.84Extended Thinking ｜ Tools | 88.60Extended Thinking ｜ Tools |
Text Arena (Coding) 编程与软件工程 | 1555.35Standard Mode | 1545.05Standard Mode |
WeirdML v2 编程与软件工程 | 77.95Thinking Level · High ｜ Tools | 82.89Thinking Level · Extra High ｜ Tools |
GPQA Diamond 科学与综合推理 | 91.31Extended Thinking | 93.60Thinking Level · High |
26.83Thinking Level · High | 56.10Thinking Level · High | |
FrontierMath v2 数学推理 | 65.96Thinking Level · High | 80.00Thinking Level · High |
SimpleBench 常识推理 | 67.60Standard Mode | 64.80Standard Mode |
BrowseComp AI Agent - 信息收集 | 84.00Thinking Enabled ｜ Tools | 84.30Thinking Level · High ｜ Tools |
MCP-Atlas AI Agent - 工具使用 | 76.80Thinking Level · High ｜ Tools | 82.20Thinking Level · High ｜ Tools |
OSWorld-Verified AI Agent - 工具使用 | 72.70Extended Thinking ｜ Tools | 83.40Extended Thinking ｜ Tools |
GDPval-AA 生产力知识 | 1606.00Extended Thinking ｜ Tools | 1890.00Extended Thinking ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Claude Opus 4.6Anthropic | Claude Opus 4.8Anthropic |
|---|---|---|
Core specsRelease | 2026-02-05 | 2026-05-28 |
Context length | 1000K | 1M |
Max output | 65536 | 128000 |
MoE | No | No |
LicenseCode Open Source | Closed Source | Closed Source |
Weights Open Source | Closed Source | Closed Source |
Commercial use | 不开源 | 不开源 |
Local deploymentWeight size | 0B | Not provided |
VRAM for weights | Not provided | Not provided |
Modality supportText Input/Output | / | / |
Image Input/Output | / | / |
ResourcesPaper / report | Introducing Claude Opus 4.6 | Introducing Claude Opus 4.8 |
DataLearner blog | Not provided | Anthropic发布Claude Opus 4.8：定价不变，编程与智能体能力小幅提升， |
