datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
minimax-m3-claude-code-tracesThis dataset was generated using teich by TeichAI
Prepare these datasets for supervised fine-tuning in just a few lines of code — see the Conversion section below.
Minimax M3 Claude Code Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by minimax/minimax-m3.
JSONL files: 31
Format
Each file is newline-delimited JSON representing a single captured agent session.
The trace schema is designed for… See the full description on the dataset page: https://huggingface.co/datasets/armand0e/minimax-m3-claude-code-traces.OctoCodingBench
OctoCodingBench: Instruction-Following Benchmark for Coding Agents
English | 中文
🌟 Overview
OctoCodingBench benchmarks scaffold-aware instruction following in repository-grounded agentic coding.
Why OctoCodingBench?
Existing benchmarks (SWE-bench, etc.) focus on task completion — whether the agent produces correct code. However, they miss a critical dimension: does the agent follow the rules while solving the task?
In real-world agentic coding, agents must… See the full description on the dataset page: https://huggingface.co/datasets/MiniMaxAI/OctoCodingBench.minimax-m3-deepsearchqa-skill-eval
MiniMax M3 DeepSearchQA Skill Eval
Evaluates minimax/minimax-m3 on google/deepsearchqa using a Pi agent, You.com MCP tools, and a research skill optimized for this harness, model, and tool surface.
MiniMax M3 Medium Reasoning with the You.com research skill reached 74.85% adjusted F1 on DeepSearchQA, above the paper's GPT-5 High Reasoning F1 result. Public artifacts are available for inspection and reproduction.
Links
GitHub:… See the full description on the dataset page: https://huggingface.co/datasets/youdotcom/minimax-m3-deepsearchqa-skill-eval.minimax-m2.7-agent
Agentic Training Traces
This directory contains raw agent trace files generated by agentic-datagen.
All assistant responses were generated by minimax/minimax-m2.7.
Trace files: 20
Training-ready tools
Use this tools payload when rendering converted examples through your training chat template.
The same structure is emitted on each converted example as the tools field.
[
{
"type": "function",
"function": {
"name": "bash",
"parameters": {… See the full description on the dataset page: https://huggingface.co/datasets/armand0e/minimax-m2.7-agent.minimax-h3-video-prompts
MiniMax H3 Video Prompts
A small, curated collection of 50 structured prompts for text-to-video and image-to-video workflows. It covers cinematic scenes, characters, animation, nature, architecture, product shots, food, social video, and fantasy environments.
Use the prompts to create video
Copy a prompt from the dataset, adapt it to your idea, then generate the finished video online.
Create an AI video with MiniMax3.org →
Dataset details… See the full description on the dataset page: https://huggingface.co/datasets/jayseanbrambila/minimax-h3-video-prompts.minimax-h3-prompt-dataset
MiniMax H3 Structured Video Prompts
A community dataset of 300 structured MiniMax H3 video prompts for text-to-video and image-to-video workflows, covering product advertising, e-commerce, fashion, beauty, food, automotive, SaaS, social media, camera motion, lighting, and commercial-use scenarios.
Each record separates reusable prompting elements such as subject, action, environment, camera motion, lighting, visual style, workflow, use case, and commercial intent.… See the full description on the dataset page: https://huggingface.co/datasets/jayseanbrambila/minimax-h3-prompt-dataset.MiniMax-M3-150k-Mixed
m3-alldomains-verified-107k
Verified distillation traces generated with faststill v0.0.1 — a pipeline that generates (prompt, reasoning, output) triplets from any OpenAI-compatible chat-completions endpoint and deterministically verifies every row before keeping it. A row is verified=true only when a machine check (executed unit tests, exact / normalized answer compare) confirmed it, so wrong labels are filtered out instead of poisoning a student model.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/empero-ai/MiniMax-M3-150k-Mixed.minimax-m2-tiny-fidelity-root-v1
minimax-m2 random CPU fixture root
A root fidelity dataset in hidden form, produced by engines/tools/hf_capture.py from malaiwah/minimax-m2-tiny-random-bf16.
The cut
the final hidden state handed to lm_head -- after the text model's final norm and immediately before the head matmul -- captured as the head module's input via torch.nn.Module.register_forward_pre_hook; replay applies the head ONLY (no final norm at replay time: the capture already sits after it).… See the full description on the dataset page: https://huggingface.co/datasets/malaiwah/minimax-m2-tiny-fidelity-root-v1.swebench-minimax-traces
swebench-minimax-traces
Complete HTTP-level agentic traces from running swebench benchmark tasks through an instrumented reverse proxy.
Each trace captures full request/response pairs including system prompts, user messages, assistant responses, tool calls and results, and token usage metadata.
Stats
Total sessions: 500
Multi-turn sessions (2+ LLM calls): 357
Total records: 22424
Total LLM requests: 11249
Format
Raw JSONL traces from the instrumented proxy.… See the full description on the dataset page: https://huggingface.co/datasets/sammshen/swebench-minimax-traces.minimax-m3-tiny-fidelity-root-v1
minimax-m3 random CPU fixture root
A root fidelity dataset in hidden form, produced by engines/tools/hf_capture.py from malaiwah/minimax-m3-tiny-random-bf16.
The cut
the final hidden state handed to lm_head -- after the text model's final norm and immediately before the head matmul -- captured as the head module's input via torch.nn.Module.register_forward_pre_hook; replay applies the head ONLY (no final norm at replay time: the capture already sits after it).… See the full description on the dataset page: https://huggingface.co/datasets/malaiwah/minimax-m3-tiny-fidelity-root-v1.intercode-minimax-traces
intercode-minimax-traces
Complete HTTP-level agentic traces from running intercode benchmark tasks through an instrumented reverse proxy.
Each trace captures full request/response pairs including system prompts, user messages, assistant responses, tool calls and results, and token usage metadata.
Stats
Total sessions: 338
Multi-turn sessions (2+ LLM calls): 299
Total records: 5838
Total LLM requests: 2919
Format
Raw JSONL traces from the instrumented proxy.… See the full description on the dataset page: https://huggingface.co/datasets/sammshen/intercode-minimax-traces.MiniMax-M2.1-Code-SFT
MiniMax M2.1 Code SFT
200 of the prompts for this dataset were sourced from MiniMaxAI/VIBE. The rest were generated.
Each prompt was given to MiniMax M2.1 with the follow tools and system prompt:
read_file - Read file contents from workspace
write_file - Write content to a file
edit_file - Replace text in a file
list_directory - List files and directories
search_code - Search for patterns in files
run_command - Execute shell commands (with timeout)
web_search - Web search (powered… See the full description on the dataset page: https://huggingface.co/datasets/TeichAI/MiniMax-M2.1-Code-SFT.minimax-h3-commercial-prompt-benchmark
MiniMax H3 Commercial Prompt Benchmark
A community benchmark evaluating MiniMax H3 video-generation prompts across commercial, cinematic and motion-focused workflows.
The benchmark tracks expected behavior, observed results, strengths, limitations and common failure modes instead of publishing prompt examples alone.
Try MiniMax H3 Text to Video →
What this dataset is
This first release contains 30 linked benchmark cases covering product advertising, ecommerce… See the full description on the dataset page: https://huggingface.co/datasets/jayseanbrambila/minimax-h3-commercial-prompt-benchmark.minimax-m3-150k-mixed
m3-alldomains-verified-107k
Verified distillation traces generated with faststill v0.0.1 — a pipeline that generates (prompt, reasoning, output) triplets from any OpenAI-compatible chat-completions endpoint and deterministically verifies every row before keeping it. A row is verified=true only when a machine check (executed unit tests, exact / normalized answer compare) confirmed it, so wrong labels are filtered out instead of poisoning a student model.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/ansulev/minimax-m3-150k-mixed.MiniMax-M2.1-8800x
MiniMax M2.1 - 8,800x
This is a reasoning dataset created using MiniMax M2.1 with reasoning effort set to high (not sure if that flag does anything for this model though).
The dataset is meant for creating distilled versions of MiniMax M2.1 by fine-tuning already existing open-source LLMs.
The prompt for this dataset cover these domains:
Science (high-school to graduate level)
Programming (general)
Web-Design
Multi-lingual creative writing
Marketing
Law
Health
Finance
Less common… See the full description on the dataset page: https://huggingface.co/datasets/TeichAI/MiniMax-M2.1-8800x.minimax-m2.1-reap-observations
[!TIP]
Support this work: donate.sybilsolutions.ai
REAP surfaces: GLM | MiniMax | Qwen | Gemma | Paper | Code | PR17 | Cerebras Collection
MiniMax-M2.1 REAP Stress Test Observations
Comprehensive stress test results for MiniMax-M2.1 models pruned with REAP (Router-weighted Expert Activation Pruning) at various compression ratios.
Dataset Description
This dataset contains 96 stress test results across 4 pruned MiniMax-M2.1 models, testing for repetition loops at… See the full description on the dataset page: https://huggingface.co/datasets/0xSero/minimax-m2.1-reap-observations.mint-minimax-traces
mint-minimax-traces
Complete HTTP-level agentic traces from running mint benchmark tasks through an instrumented reverse proxy.
Each trace captures full request/response pairs including system prompts, user messages, assistant responses, tool calls and results, and token usage metadata.
Stats
Total sessions: 242
Multi-turn sessions (2+ LLM calls): 145
Total records: 1388
Total LLM requests: 694
Format
Raw JSONL traces from the instrumented proxy. Each line is… See the full description on the dataset page: https://huggingface.co/datasets/sammshen/mint-minimax-traces.MiniMax_M2.7_Distilled_5k
MiniMax-M2.7 Thinking Distilled Dataset
A 5,000-example synthetic reasoning dataset mirroring MiniMax-M2.7 Thinking interleaved reasoning style, with <think> tags separating reasoning steps from final responses.
Dataset
File: minimax_m2.7_distilled_5k.jsonl (5,000 lines, ~3.5 MB)
Each example is a JSON object with:
Field
Type
Description
instruction
str
The user query / task prompt
thinking
str
Interleaved reasoning trace wrapped in <think> tags… See the full description on the dataset page: https://huggingface.co/datasets/WithinUsAI/MiniMax_M2.7_Distilled_5k.repro-understanding-sam-through-minimax-perspective-traces
Agent traces
Agent sessions published from a Trackio Logbook.
swebench-minimax-traces-2
swebench-minimax-traces-2
Complete HTTP-level agentic traces from running swebench benchmark tasks through an instrumented reverse proxy.
Each trace captures full request/response pairs including system prompts, user messages, assistant responses, tool calls and results, and token usage metadata.
Stats
Total sessions: 250
Multi-turn sessions (2+ LLM calls): 210
Total records: 15198
Total LLM requests: 7623
Format
Raw JSONL traces from the instrumented proxy.… See the full description on the dataset page: https://huggingface.co/datasets/sammshen/swebench-minimax-traces-2.minimax-m2.1-1000x
MiniMax M2.1
This is a reasoning dataset created using MiniMax M2.1 with a reasoning depth set to high.
The dataset is meant for creating distilled versions of MiniMax M2.1 by fine-tuning already existing open-source LLMs.
Some of these prompts are from reedmayhew, the rest were generated
Stats:
Cost: $ 3.83 USD
Total Tokens (Input + Output): 3.28 M
minimax-m3-repitition
MiniMax M3 Repetition Samples
Reviewed samples from z-lab/minimax-m3-regen whose regenerated assistant responses contain repeated blocks, repeated tails, or punctuation loops.
Summary
Rows: 3,799
Source counts:
nemotron: 2,704
evol_codealpaca: 464
opencodeinstruct: 473
openr1_math: 158
Severity counts:
high: 938
medium: 2,079
low: 782
The JSONL rows contain the original dataset row plus _repetition and _repetition_review metadata.
error_recovery_gemma4x2_minimax.jsonlminimax-2.7-Bug-Prominimax_run3_equiv_partial_pyramid-panguMiniMax-SFTminimax-m2-250xmodal-vllm-cache-h200-minimax-v43repro-minimax-learning-of-interpretable-factored-stochastic-policies-from-conjoint-data-with-unc
Reproduction: Learning Interpretable Factored Policies from Conjoint Data with MiniMax Learning
Paper Information
Title: Learning Interpretable Factored Policies from Conjoint Data with Uncertainty-Calibrated Minimax Learning
OpenReview ID: GJblFvJcMb
Conference: ICML 2026
Task: Optimal treatment selection from conjoint survey data using minimax game-theoretic framework
Reproduction Summary
This reproduction evaluates the paper's claims about… See the full description on the dataset page: https://huggingface.co/datasets/sabaridsnfuji/repro-minimax-learning-of-interpretable-factored-stochastic-policies-from-conjoint-data-with-unc.sft_ablations_scientific_minimax_v1