免费 / 开源由 DeepSeek-V4 与 R1 驱动的高效开源权重与托管推理 AI,token 成本极低。
Aide Dev 记录了 5,079 月访问量,月度环比增长为 -1.9%,定价为付费。Imandra 记录了 1,466 月访问量,增长率为 -25.7%,定价为免费增值。

Imandra is a neurosymbolic reasoning platform that turns source code and system logic into precise mathematical models, so AI agents and engineers can prove properties, find edge-case bugs, and verify behavior before it ships. Its flagship CodeLogician tool plugs into AI coding assistants such as Codex, Cursor, and Gemini, while ImandraX powers automated reasoning for financial markets, autonomous systems, and regulated industries.
| 维度 / 指标 | ![]() | |
|---|---|---|
| 官方域名 | aide.dev | imandra.ai |
| 定价模式 | 付费 | 免费增值 |
| 月访问量 | 5079 (5,079) | 1466 (1,466) |
| 月度环比流量变化 | -98 (-1.9%) | -508 (-25.7%) |
| 主要市场与份额 | 🇺🇸 美国 (79.7%) | 🇮🇳 印度 (100%) |
| 直接访问 / 月 | 2017 | 567 |
| 用户评分 | 4.5 | 4.4 |
| 支持的平台 | 网站 | 网站 |
面向软件工程的 AI 智能体,在推理阶段进行扩展
智能体在测试阶段投入额外算力进行推理、探索与验证,而不是仅依赖一次前向推理就给出修复。
面向真实代码仓库的任务:定位相关代码、理解意图,并产出能够解决问题的改动。
团队公布实测成绩,包括在 SWE-bench Verified 上报告的 62.2% 解决率。
持续发布工程笔记,覆盖智能体架构、评测方法与构建编码系统的经验教训。

Imandra is a neurosymbolic reasoning platform that turns source code and system logic into precise mathematical models, so AI agents and engineers can prove properties, find edge-case bugs, and verify behavior before it ships. Its flagship CodeLogician tool plugs into AI coding assistants such as Codex, Cursor, and Gemini, while ImandraX powers automated reasoning for financial markets, autonomous systems, and regulated industries.
Imandra's reasoning engine builds formal models of programs and autonomous systems, then exhaustively analyzes behavior to surface edge cases that testing and statistical models miss.
CodeLogician translates a codebase into mathematical logic and assembles a MetaModel spanning every file and dependency, giving AI assistants a complete structural context with no learning phase.
Instead of sampling inputs, Imandra decomposes state space into behavioral regions so decisions and boundary conditions are examined exhaustively and reproducibly.
Benchmarks published by Imandra show that augmenting LLMs with formal reasoning through CodeLogician closed a 41-47 percentage point accuracy gap on software-analysis reasoning tasks compared with LLM-only approaches.
免费 / 开源由 DeepSeek-V4 与 R1 驱动的高效开源权重与托管推理 AI,token 成本极低。
会学习你编码品味的 AI 编程智能体
发现并自动修复漏洞的一体化应用安全平台
覆盖整个软件开发周期的智能体编排平台