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。
Gemini 2.5-Pro
Google Deep Mind
Each axis is a category average, normalized to a 100-point radar.
Relative edge: none clear / Relative gap: Agent能力评测 -23.2
Relative edge: Agent能力评测 +23.2 / 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
GPT-5 · 119.28
Best single
GPT-5 · CodeClash 1360.00
Modality coverage
Gemini 2.5-Pro · 2 modalities
Head to head
20
Benchmarks
2
Wins
18
Losses
-23.59
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.
20 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Gemini 2.5-Pro | GPT-5 |
|---|---|---|
ARC-AGI 综合评估 | 37.00Thinking Enabled | 65.70Thinking Level · High |
ARC-AGI-2 综合评估 | 4.90Thinking Enabled | 9.90Thinking Level · High |
GPQA Diamond 综合评估 | 86.40Thinking Enabled | 87.30Thinking Enabled ｜ Tools |
HLE 综合评估 | 21.60Thinking Enabled | 35.20Thinking Enabled ｜ Tools |
CodeClash 编程与软件工程 | 1125.00Standard Mode ｜ Tools | 1360.00Standard Mode ｜ Tools |
SWE-bench Verified 编程与软件工程 | 67.20Thinking Enabled | 72.80Thinking Level · High |
AIME2025 数学推理 | 88.00Thinking Enabled | 99.60Thinking Enabled ｜ Tools |
FrontierMath 数学推理 | 11.00Standard Mode | 26.30Thinking Level · High ｜ Tools |
2.10Standard Mode | 12.50Thinking Level · High | |
IMO 2024 数学推理 | 19.00Thinking Enabled | 11.00Thinking Enabled |
IMO 2025 数学推理 | 15.20Thinking Enabled | 29.00Thinking Enabled |
IMO-ProofBench 数学推理 | 55.20Thinking Enabled | 59.00Thinking Enabled |
17.60Thinking Enabled | 20.00Thinking Enabled | |
Terminal-Bench AI Agent - 工具使用 | 25.30Thinking Enabled | 43.80Thinking Enabled ｜ Tools |
MMMU 多模态理解 | 82.00Thinking Enabled | 84.20Thinking Level · High |
Simple Bench 常识推理 | 62.40Thinking Enabled | 56.70Thinking Level · High |
Aider-Polyglot Agent能力评测 | 83.1032K | 88.00Thinking Level · High |
τ²-Bench - Telecom Agent能力评测 | 54.00Thinking Enabled ｜ Tools | 96.70Thinking Level · High ｜ Tools |
IF Bench 指令跟随 | 49.00Thinking Enabled ｜ Tools | 73.10Thinking Level · High |
BrowseComp AI Agent - 信息收集 | 7.80Thinking Enabled ｜ Tools | 54.90Thinking Enabled ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Gemini 2.5-ProGoogle Deep Mind | GPT-5OpenAI |
|---|---|---|
Core specsRelease | 2025-06-05 | 2025-08-07 |
Context length | 1000K | 400K |
Max output | 65536 | 131072 |
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 | / | / |
ResourcesPaper / report | Try the latest Gemini 2.5 Pro before general availability. | Introducing GPT-5 |
DataLearner blog | Google发布Gemini 2.5 Pro: Gemini系列第一个2.5版本的模型，最高支持200万上下文，全模态输入，推理大模型，LMArena排名第一 | OpenAI发布GPT-5：这是一个包含实时路由的AI系统，而不仅仅是一个模型 |
GPT-5
OpenAI
