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
Each axis is a category average, normalized to a 100-point radar.
Relative edge: 常识推理 +5.9 / Relative gap: 编程与软件工程 -6.8
Relative edge: 编程与软件工程 +6.8 / Relative gap: 常识推理 -5.9
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
Opus 4.7 · 72.03
Best single
Opus 4.7 · GPQA Diamond 94.20
Modality coverage
Claude Opus 4.6 · 2 modalities
Head to head
14
Benchmarks
2
Wins
11
Losses
-1.96
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.
14 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Claude Opus 4.6 | Opus 4.7 |
|---|---|---|
ARC-AGI 综合评估 | 92.00Extended Thinking | 93.50Thinking Level · High |
ARC-AGI-2 综合评估 | 66.30Extended Thinking | 75.80Thinking Level · High |
GPQA Diamond 综合评估 | 91.31Extended Thinking | 94.20Extended Thinking |
HLE 综合评估 | 53.00Extended Thinking ｜ Tools | 54.70Extended Thinking ｜ Tools |
LiveBench 综合评估 | 76.33Thinking Level · High | 76.91Deep Thinking Mode |
MMLU 综合评估 | 91.05Extended Thinking | 91.50Standard Mode |
SWE-bench Verified 编程与软件工程 | 80.84Extended Thinking ｜ Tools | 87.60Extended Thinking ｜ Tools |
FrontierMath 数学推理 | 40.70Thinking Level · High | 43.80Thinking Level · Extra High |
22.90Thinking Level · High | 22.90Thinking Level · Extra High | |
Simple Bench 常识推理 | 67.60Standard Mode | 61.70Standard Mode |
BrowseComp AI Agent - 信息收集 | 84.00Thinking Enabled ｜ Tools | 79.30Extended Thinking ｜ Tools |
MCP-Atlas AI Agent - 工具使用 | 76.80Deep Thinking Mode ｜ Tools | 79.10Deep Thinking Mode ｜ Tools |
OSWorld-Verified AI Agent - 工具使用 | 72.70Extended Thinking ｜ Tools | 78.00Extended Thinking ｜ Tools |
Terminal Bench 2.0 AI Agent - 工具使用 | 65.40Extended Thinking ｜ Tools | 69.40Extended Thinking ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Claude Opus 4.6Anthropic | Opus 4.7Anthropic |
|---|---|---|
Core specsRelease | 2026-02-05 | 2026-04-16 |
Context length | 1000K | 1000K |
Max output | 65536 | 131072 |
MoE | No | No |
LicenseCode Open Source | Not provided | Not provided |
Weights Open Source | Not provided | Not provided |
Commercial use | 不开源 | 不开源 |
Modality supportText Input/Output | / | / |
Image Input/Output | / | / |
ResourcesPaper / report | Introducing Claude Opus 4.6 | Introducing Claude Opus 4.7 |
DataLearner blog | Not provided | Anthropic发布Claude Opus 4.7：编程能力大幅跃升，视觉分辨率提升超3倍，首个搭载网络安全防护机制的旗舰模型！ |
Opus 4.7
Anthropic
