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 Sonnet 4
Anthropic
Each axis is a category average, normalized to a 100-point radar.
Relative edge: none clear / Relative gap: 数学推理 -33.0
Relative edge: 数学推理 +33.0 / Relative gap: none clear
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
Gemini 3.0 Flash · 70.92
Best single
Gemini 3.0 Flash · AIME2025 99.70
Modality coverage
Gemini 3.0 Flash · 4 modalities
Head to head
11
Benchmarks
1
Wins
10
Losses
-14.85
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.
11 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Claude Sonnet 4 | Gemini 3.0 Flash |
|---|---|---|
ARC-AGI-2 综合评估 | 5.90Thinking Enabled | 33.60Thinking Enabled |
HLE 综合评估 | 9.60Thinking Enabled | 43.50Thinking Enabled ｜ Tools |
LiveBench 综合评估 | 61.2764K | 72.40Thinking Level · High |
GPQA Diamond 科学与综合推理 | 83.80Deep Thinking Mode ｜ Tools | 90.40Thinking Enabled |
SWE-Bench Pro - Public 编程与软件工程 | 42.70Thinking Enabled | 49.60Thinking Level · High ｜ Tools |
SWE-bench Verified 编程与软件工程 | 72.70Thinking Enabled ｜ Tools | 68.70Thinking Enabled |
AIME2025 数学推理 | 85.00Deep Thinking Mode ｜ Tools | 99.70Thinking Enabled ｜ Tools |
SimpleBench 常识推理 | 45.50Thinking Enabled | 61.10Standard Mode |
τ²-Bench Agent能力评测 | 52.00Standard Mode ｜ Tools | 90.20Thinking Enabled ｜ Tools |
Claw Bench OpenClaw智能体能力综合测评 | 77.80Thinking Enabled ｜ Tools | 85.70Thinking Enabled ｜ Tools |
Pinch Bench OpenClaw智能体能力综合测评 | 80.50Thinking Enabled ｜ Tools | 85.20Thinking Enabled ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Claude Sonnet 4Anthropic | Gemini 3.0 FlashGoogle Deep Mind |
|---|---|---|
Core specsRelease | 2025-05-23 | 2025-12-17 |
Context length | 200K | 2000K |
Max output | 64000 | 65536 |
MoE | No | No |
Supported modes | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode）深度思考（Deeper Thinking Mode） | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode）深度思考（Deeper Thinking Mode） |
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 | / | / |
Audio Input/Output | Not provided | / |
Video Input/Output | Not provided | / |
ResourcesPaper / report | Introducing Claude 4 | Gemini 3 Flash: frontier intelligence built for speed |
DataLearner blog | Anthropic发布Claude4，全球最强编程大模型，大幅提升AI Agent系统所需的各项能力，最长可以7小时连续工作，80%工程任务自主完成 | Gemini 3 Flash：Google 在 12 月 17 日发布的新一代默认模型 |
Gemini 3.0 Flash
Google Deep Mind
