MiniMax M3.1 Flash

Leaderboards

Where MiniMax M3.1 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 MiniMax M3.1 Flash has a published result on, out of 100%, with its rank among the models that took it.

• CharXiv-R

  #20 / 60

  Reading and reasoning about charts from scientific papers.

  82.5%

  Self-reported

• DeepSWE 1.1

  #13 / 50

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

  73.7%

  Self-reported · In the coding score

• Terminal-Bench 4.0

  #14 / 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.

  38.1%

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

• NL2Repo

  #4 / 25

  NL2Repo evaluates long-horizon coding capabilities including repository-level understanding, where models must generate or modify code across entire repositories from natural language specifications.

  61.2%

  Self-reported · In the coding score

• Agents' Last Exam

  #16 / 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.

  29.6%

  Self-reported · In the agents score

• CyberGym

  #9 / 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.

  82.2%

  Self-reported

• OSWorld 2.0

  #10 / 17

  OSWorld 2.0 is a benchmark of 108 long-horizon, real-world computer-use workflows spanning everyday and professional tasks. Each task is an end-to-end workflow that takes human users a median of about 1.6 hours, scored with a binary-completion metric, and targets challenges such as dynamic environments, cross-source reasoning, and implicit-state inference.

  61.7%

  Self-reported

• BabyVision

  #2 / 16

  A benchmark for early-stage visual reasoning and perception on child-like vision tasks.

  91.0%

  Self-reported

• GDP.pdf

  #13 / 14

  GDP.pdf is a knowledge-work vision benchmark that evaluates models on economically valuable professional tasks presented as visual documents (PDFs), testing document-based reasoning, chart and table interpretation, and problem solving without tools.

  20.8%

  Self-reported

• Humanity's Last Exam (no tools, text-only)

  #7 / 14

  Text-only Humanity's Last Exam variant evaluated without tool use.

  43.4%

  Self-reported

• PostTrainBench

  #3 / 9

  PostTrainBench evaluates a model's ability to autonomously post-train base models. Given pretrain-only base models, the agent must complete the full pipeline of data synthesis, training, evaluation, and iteration within a time budget, scored across downstream benchmarks such as AIME2025, BFCL, GPQA Main, GSM8K, and HumanEval.

  39.3%

  Self-reported

• AutomationBench v1.0.6

  #1 / 6

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

  62.2%

  Self-reported

• BankerToolBench

  #2 / 3

  BankerToolBench is a public benchmark that evaluates models on banking and finance tool-use tasks. Models are scored against dataset rubrics, measuring their ability to correctly invoke tools and complete multi-step financial workflows.

  76.2%

  Self-reported

• LOCA-Bench (256k)

  #1 / 2

  LOCA-Bench is a long-context agentic benchmark. The 256k variant evaluates agents using the official ReAct mode with an environment description length of 256k tokens, measuring how well models reason and act over very long contexts.

  60.3%

  Unverified