See key specs and per-benchmark scores for each model/mode. Scroll horizontally for all columns. 当前对比 3 个模型的评测数据与核心参数。
Step 3.5 Flash
StepFunAI
Each axis is a category average, normalized to a 100-point radar.
Relative edge: AI Agent - 信息收集 +8.4 / Relative gap: 综合评估 -10.3
Relative edge: 综合评估 +10.3 / Relative gap: AI Agent - 信息收集 -8.4
Relative edge: none clear / Relative gap: 数学推理 -6.7
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
Qwen3-Max-Thinking · 81.50
Best single
Step 3.5 Flash · AIME2025 99.80
Modality coverage
Kimi K2.5 · 2 modalities
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.
10 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Step 3.5 Flash | Kimi K2.5 | Qwen3-Max-Thinking |
|---|---|---|---|
ARC-AGI 综合评估 | 56.50Thinking Enabled ｜ Tools | 65.30Thinking Enabled | -- |
LiveCodeBench 编程与软件工程 | 86.40Thinking Enabled | 85.00Thinking Enabled | 85.90Thinking Enabled |
SWE-bench Verified 编程与软件工程 | 74.40Thinking Enabled | 76.80Thinking Enabled ｜ Tools | 75.30Thinking Enabled |
AIME2025 数学推理 | 99.80Thinking Enabled ｜ Tools | 96.10Thinking Enabled | -- |
IMO-AnswerBench 数学推理 | 86.70Thinking Enabled ｜ Tools | 81.80Thinking Enabled | 83.90Thinking Enabled |
τ²-Bench Agent能力评测 | 88.20Thinking Enabled ｜ Tools | -- | 82.10Thinking Enabled ｜ Tools |
BrowseComp AI Agent - 信息收集 | 69.00Thinking Enabled ｜ Tools | 60.60Thinking Enabled ｜ Tools | -- |
Terminal Bench 2.0 AI Agent - 工具使用 | 51.00Thinking Enabled ｜ Tools | 50.80Thinking Enabled ｜ Tools | -- |
Claw Bench OpenClaw智能体能力综合测评 | 84.90Thinking Enabled ｜ Tools | 81.70Thinking Enabled ｜ Tools | -- |
Pinch Bench OpenClaw智能体能力综合测评 | 85.30Thinking Enabled ｜ Tools | 84.80Thinking Enabled ｜ Tools | 80.30Thinking Enabled ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Step 3.5 FlashStepFunAI | Kimi K2.5Moonshot AI | Qwen3-Max-Thinking阿里巴巴 |
|---|---|---|---|
Core specsRelease | 2026-02-02 | 2026-01-27 | 2026-01-26 |
Context length | 256K | 256K | 1000K |
Parameters | 1960 | 10000 | 10000 |
Active parameters | 110 | 320 | Not provided |
Max output | 16384 | 16384 | 32768 |
MoE | Yes | Yes | Yes |
Supported modes | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode） | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode） | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode） |
LicenseCode Open Source | Not provided | Not provided | Not provided |
Weights Open Source | Not provided | Not provided | Not provided |
Commercial use | 免费商用授权 | 免费商用授权 | 不开源 |
Modality supportText Input/Output | / | / | / |
Image Input/Output | / | / | / |
Audio Input/Output | / | / | / |
Video Input/Output | / | / | / |
Embedding Input/Output | / | / | / |
ResourcesPaper / report | Step 3.5 Flash | Kimi K2.5: Visual Agentic Intelligence | Qwen3-Max-Thinking: Pushing the Limits of Reasoning via Test-Time Scaling |
DataLearner blog | 阶跃星辰StepFun开源激活参数只有11B的Step-3.5-Flash模型，但是评测结果和Kimi K2.5、Qwen3-Max-Thinking差不多，最高推理速度可以达到350tokens/s！ | 重磅！Kimi K2.5发布，依然免费开源！原生多模态MoE架构，全球最大规模参数的开源模型之一，官方评测结果比肩诸多闭源模型！可以驱动100个子Agent执行！ | Not provided |
Kimi K2.5
Moonshot AI
Qwen3-Max-Thinking
阿里巴巴
