加载中...
加载中...
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
Kimi K2.5
Qwen3-Max-Thinking
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
Performance benchmarks
Compare benchmark results across thinking modes and tool usage.
Best Overall
Qwen3-Max-Thinking · 83.90
Best Single
Step 3.5 Flash · AIME2025 99.80
Thinking Mode (Default)
Step 3.5 Flash · 1 All Modes
Higher is usually better; “—” means no score.
Complete scores for each model/mode across selected benchmarks.
Higher is usually better; “—” means no score.
| Benchmark | ST Step 3.5 FlashStepFunAI | KI Kimi K2.5Moonshot AI | QW Qwen3-Max-Thinking阿里巴巴 |
|---|---|---|---|
| 数学推理 | |||
AIME2025 | 99.80 | 96.10 | — |
IMO-AnswerBench | 86.70 | 81.80 | 83.90 |
| AI Agent - 信息收集 | |||
BrowseComp | 69.00 | 60.60 | — |
| AI Agent - 工具使用 | |||
Terminal Bench 2.0 | 51.00 | 50.80 | — |
Feature compare
Licensing, MoE architecture, and multi-modality support.
| Features & specs | ST Step 3.5 FlashStepFunAI | KI Kimi K2.5Moonshot AI | QW Qwen3-Max-Thinking阿里巴巴 |
|---|---|---|---|
Model snapshots | |||
Organization | StepFunAI | Moonshot AI | 阿里巴巴 |
模型全名 | Step 3.5 Flash | Kimi K2.5 | Qwen3-Max-Thinking |
模型简介 | Not provided | Not provided | Not provided |
模型类型 | 聊天大模型 | 多模态大模型 | 推理大模型 |
模型代号 | stepfun-flash-3-5 | kimi-k2-5 | qwen3-max |
Release | 2026-02-02 | 2026-01-27 | 2026-01-26 |
MoE | Yes | Yes | Yes |
规格与性能 | |||
Context length | 256K | 256K | 1000K |
Parameters | 1960 | 10000 | 10000 |
激活参数量 | 110 | 320 | Not provided |
模型规模 | 100b | 100b | 100b |
模型大小 | 399GB | 595GB | Not provided |
推理速度 | |||
推理等级 | |||
最大输出 | 16384 | 16384 | 32768 |
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） |
开源与许可 | |||
Code 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 support | |||
Text Input/Output | / | / | / |
Image Input/Output | / | / | / |
Audio Input/Output | / | / | / |
Video Input/Output | / | / | / |
Embedding Input/Output | / | / | / |
API 接口详情 | |||
Text 价格 | Input: 0.0Output: 0.0Cache: 0.0Input (Extended): 0.0 | Input: 0.6 美元/100 万tokensOutput: 3 美元/100 万tokensCache: 0.1 美元/100 万tokens | Input: 1.2 美元/100万 tokensOutput: 6 美元/100万 tokensInput (Extended): 2.4 美元/100万 tokensOutput (Extended): 12 美元/100万 tokensThreshold: 32K |
Image API pricing | Not provided | Input: 0.6 美元/100 万tokensCache: 0.1 美元/100 万tokens | Not provided |
Audio API pricing | Not provided | Not provided | Not provided |
Video API pricing | Not provided | Not provided | Not provided |
Embedding API pricing | Not provided | Not provided | Not provided |
Resources | |||
GitHub | Repo | Repo | Not provided |
Hugging Face | Model Page | Model Page | Not provided |
Official Page | Not provided | Not provided | Not provided |
Guides | Not provided | Not provided | Not provided |
Papers | Step 3.5 Flash | Kimi K2.5: Visual Agentic Intelligence | Qwen3-Max-Thinking: Pushing the Limits of Reasoning via Test-Time Scaling |
DataLearnerAI | 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 |
API pricing
Side-by-side input/output token pricing