MiMo-V2.6-Flash

Leaderboards

Where MiMo-V2.6-Flash sits among the top 20 in each category.

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Test results

Every test MiMo-V2.6-Flash has a published result on, out of 100%, with its rank among the models that took it.

• DeepSWE 1.1

  #24 / 50

  DeepSWE 1.1 evaluates software engineering agents using the mini-swe-agent harness.

  67.9%

  Self-reported · In the coding score

• Terminal-Bench 2.1

  #9 / 46

  Terminal-Bench 2.1 is an updated release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal. It evaluates how well models handle real-world, end-to-end tasks autonomously, including compiling code, training models, setting up servers, system administration, data science workflows, and security tasks.

  87.6%

  Self-reported

• OSWorld-Verified

  #8 / 28

  OSWorld-Verified is a verified subset of OSWorld, a scalable real computer environment for multimodal agents supporting task setup, execution-based evaluation, and interactive learning across Ubuntu, Windows, and macOS.

  80.8%

  Self-reported

• Terminal-Bench 4.0

  #19 / 28

  Terminal-Bench 4.0 evaluates AI agents on real-world work in terminal and command-line containerized environments across 66 tasks, with emphasis on science-adjacent and frontier engineering problems. Relative to earlier Terminal-Bench releases, 4.0 increases timeouts and adaptively raises RAM/CPU on selected tasks to reduce harness and resource confounds.

  28.8%

  Self-reported · In the coding score· In the agents score

• Agents' Last Exam

  #18 / 24

  Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.

  27.6%

  Self-reported · In the agents score

• CyberGym

  #1 / 21

  CyberGym is a benchmark for evaluating AI agents on cybersecurity tasks, testing their ability to identify vulnerabilities, perform security analysis, and complete security-related challenges in a controlled environment.

  95.1%

  Self-reported

• Program Bench

  #15 / 18

  Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation. It spans 200 tasks from small CLI tools to large systems such as FFmpeg and SQLite, with submissions judged against more than 248,000 fuzz-generated behavioral tests.

  26.0%

  Self-reported · In the coding score

• ExploitBench

  #8 / 8

  ExploitBench is a cybersecurity benchmark that evaluates a model's ability to discover and exploit software vulnerabilities, reported as the fraction of challenges where the model captures the target (Cap%).

  25.3%

  Self-reported

• ExploitGym

  #8 / 8

  ExploitGym is a large-scale, realistic benchmark built from real-world vulnerabilities across userspace programs, Google's V8 engine, and the Linux kernel. Given a vulnerability and a proof-of-vulnerability input, agents must craft a working end-to-end exploit that achieves unauthorized code execution.

  6.0%

  Self-reported

• SEC-bench Pro

  #7 / 7

  SEC-bench Pro is a self-evolving software-security benchmark that measures agent bug hunting on critical, high-complexity systems. It instantiates validated vulnerabilities across the V8 and SpiderMonkey JavaScript engines as reproducible vulnerability-discovery and proof-of-concept-generation tasks with oracle-based validation.

  47.5%

  Self-reported

• AutomationBench v1.0.6

  #3 / 6

  AutomationBench v1.0.6 evaluates agent performance on multi-step workflow automation tasks.

  52.3%

  Self-reported

• JobBench

  #6 / 6

  JobBench evaluates agents on work-related tasks, as reported in the MiMo-V2.6 release.

  61.2%

  Self-reported

• Toolathlon-Verified

  #4 / 5

  Verified Toolathlon evaluation of agent tool-use capability, as reported in the MiMo-V2.6 release.

  73.6%

  Self-reported

• MiMo Coding Bench

  #4 / 4

  MiMo Coding Bench evaluates coding-agent capabilities on software engineering tasks reported with the MiMo model family.

  61.2%

  Self-reported

• MiMo Cyber Bench

  #2 / 2

  Xiaomi's in-house cybersecurity evaluation, reported with MiMo-V2.6.

  77.2%

  Unverified

• MiMo Visual Coding

  #2 / 2

  Xiaomi's in-house evaluation of visual coding capability, reported with MiMo-V2.6.

  71.5%

  Unverified