datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
arc-agi3-kimi-k2.7-su15
ARC-AGI-3 su15 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game su15, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-su15.arc-agi3-kimi-k2.7-g50t
ARC-AGI-3 g50t — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game g50t, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-g50t.arc-agi3-kimi-k2.7-ls20
ARC-AGI-3 ls20 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game ls20, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-ls20.arc-agi3-kimi-k2.7-tr87
ARC-AGI-3 tr87 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game tr87, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-tr87.arc-agi3-kimi-k2.7-ar25
ARC-AGI-3 ar25 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game ar25, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-ar25.arc-agi3-kimi-k2.7-ft09
ARC-AGI-3 ft09 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game ft09, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-ft09.discoverphysics-kimi-k2.7-ara
DiscoverPhysics × Kimi K2.7 (kimi-code CLI, thinking=on) — 11-world ARA knowledge artifacts
Agent-Native Research Artifacts (ARA) produced by a Kimi K2.7 (kimi-code CLI, thinking=on) coding-agent session solving all
11 worlds of the DiscoverPhysics scientific-discovery
benchmark (seed 0, noise_frac 0.075, ≤16 experiment rounds), driven through the same
harness-agnostic bridge and ARA scaffold as the sibling fable run. Official verdicts:
1/11 PASS — criteria and per-world numbers… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/discoverphysics-kimi-k2.7-ara.kimi-k2.6-reap-observations-v1
Kimi-K2.6 REAP Observation Data (v1)
Per-layer expert routing + activation statistics captured from moonshotai/Kimi-K2.6
under the REAP layerwise observer (PR #17, CerebrasResearch/reap).
What this is
This dataset contains the observer output of a full REAP calibration pass on
Kimi-K2.6. It is not a pruned model. Each record describes per-token routing
decisions, expert activation norms, and the REAP saliency ingredients for every
MoE layer of the base model.
Downstream… See the full description on the dataset page: https://huggingface.co/datasets/mateowilliam/kimi-k2.6-reap-observations-v1.kimi-k2.6-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.
Kimi K2.6 Claude Code Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by moonshotai/kimi-k2.6.
JSONL files: 36
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/kimi-k2.6-claude-code-traces.kimi-k2.6-reap-observations-v1
Kimi-K2.6 REAP Observation Data (v1)
Per-layer expert routing + activation statistics captured from moonshotai/Kimi-K2.6
under the REAP layerwise observer (PR #17, CerebrasResearch/reap).
What this is
This dataset contains the observer output of a full REAP calibration pass on
Kimi-K2.6. It is not a pruned model. Each record describes per-token routing
decisions, expert activation norms, and the REAP saliency ingredients for every
MoE layer of the base model.
Downstream… See the full description on the dataset page: https://huggingface.co/datasets/0xSero/kimi-k2.6-reap-observations-v1.arc-agi3-kimi-k2.7-s5i5
ARC-AGI-3 s5i5 — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game s5i5, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-s5i5.Kimi-K2.5-Reasoning-1M-Cleaned
🪐 Kimi-K2.5-Reasoning-1M-Cleaned
Kimi-K2.5-Reasoning-1M-Cleaned is a cleaned derivative of ianncity/KIMI-K2.5-1000000x. It preserves the original four-config layout from the source dataset and rewrites each record into a unified reasoning-SFT schema with id, conversations, input, output, domain, and meta.
Summary
Source dataset: ianncity/KIMI-K2.5-1000000x
Source author: ianncity
Teacher model recorded in meta.teacher_model: KIMI-K2.5
Token lengths computed with… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/Kimi-K2.5-Reasoning-1M-Cleaned.DeepSWE-Agent-Kimi-K2-Trajectories-2.8KKIMI-K2.5-1000000x
KIMI-K2.5-1000000x
1,000,000 reasoning traces distilled from KIMI-K2.5 on high reasoning, (Each subset has different questions)
Distribution:
Coding: 50% (Includes: Webdev, Python, C++, Java, JS, C, Ruby, Lua, Rust, and C#)
Science: 20% (Physics, Chemistry, Biology) - 100k more completions in the PHD-Science subset
Math: 15% (Algebra, Calculus, Probability) - 200k more completions in kimiMath200k.jsonl
Computer Science: 5%
Logical Questions: 5%
Creative Writing: 5%… See the full description on the dataset page: https://huggingface.co/datasets/ianncity/KIMI-K2.5-1000000x.arc-agi3-kimi-k2.7-r11l
ARC-AGI-3 r11l — Agent Trajectories (kimi-k2.7)
Gameplay trajectories from the harness×model pair kimi-k2.7 playing the
ARC-AGI-3 game r11l, part of the
ARA-as-world-model generalization experiment. The agent builds a structured world model
(an Agent-Native Research Artifact) live during play and consults it to crack levels it
cannot solve from cold exploration.
One dataset repo per harness×model×game: sibling repos
arc-agi3-<harness>-<model>-<game> hold the same game played by… See the full description on the dataset page: https://huggingface.co/datasets/AgentNativeResearchLab/arc-agi3-kimi-k2.7-r11l.Creative-Writing-KimiK2.5-Cleaned
Creative-Writing-KimiK2.5-Cleaned
Cleaned creative writing SFT dataset from Kimi K2.5 (655 samples). Prompts cleaned, thinking traces preserved.
Format
Each line is a JSON object with:
messages: list of message dicts with roles (system, user, assistant)
System: writing quality instructions
User: cleaned creative writing prompt
Assistant: creative writing response (may include <think> traces)
Stats
Metric
Value
Total prompt tokens
80… See the full description on the dataset page: https://huggingface.co/datasets/Crownelius/Creative-Writing-KimiK2.5-Cleaned.DeepSWE-Agent-Kimi-K2-Trajectories-Rejection-Samplingkimi-k25-tiny-cpu-repro-v1
Kimi K2.5 / K2.6 / K2.7-Code complete tiny random BF16 CPU fixture
Untrained independently seeded random weights; no upstream weights or training data.
This is a reproducibility fixture, not useful language modeling or production quality evidence.
Runtime and lineage
Upstream moonshotai/Kimi-K2.7-Code@74797c9c62378b951a1f6fcf5c4631024e9b8bef.
Actual loaded class: Kimi_K25ForConditionalGeneration. Complete untied head and real small vision tower/projector.… See the full description on the dataset page: https://huggingface.co/datasets/malaiwah/kimi-k25-tiny-cpu-repro-v1.Kimi-K2.5-Reasoning-General-Sharded
Kimi-K2.5-Reasoning-General-Sharded
Byte-preserving sequential 100 MB JSONL shards of selected files from Jackrong/Kimi-K2.5-Reasoning-1M-Cleaned. All credit for data generation and upstream curation belongs to the source authors. See the upstream dataset card for attribution, source descriptions and license terms.
Included files: General-Distillation.jsonl.
No filtering, shuffling, normalization, tokenization or truncation was performed. Complete records and all original fields… See the full description on the dataset page: https://huggingface.co/datasets/placeholderlabs/Kimi-K2.5-Reasoning-General-Sharded.kimi-k2.6-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.
Kimi K2.6 Claude Code Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by moonshotai/kimi-k2.6.
JSONL files: 36
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/JBrightmanAI/kimi-k2.6-claude-code-traces.kimi-k2.6-agentThis 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.
Kimi K2.6 Agent Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by moonshotai/kimi-k2.6.
JSONL files: 15
Training-ready tools
A complete configured tools schema snapshot is embedded in the collapsed section at the bottom of this README.… See the full description on the dataset page: https://huggingface.co/datasets/armand0e/kimi-k2.6-agent.combined-reasoning-opus-4.6-opus-4.7-kimi-k2.5-kimi-k2.6-glm-5.1
Combined Reasoning Distill — Multi-Model
A large-scale unified reasoning dataset combining thinking and chain-of-thought traces distilled from frontier models, normalized into a single consistent schema for fine-tuning. Includes data from Claude (Opus 4.5/4.6/4.7, Sonnet 4.5/4.6, Haiku 4.5), GPT (5.1/5.2), Gemini 3 Pro Preview, Kimi (K2/K2.5/K2.6), GLM (4.6/4.7/5.1), MiniMax M2.1, Grok Code Fast 1, and more.
Schema
Every row has a single field:
Field
Type… See the full description on the dataset page: https://huggingface.co/datasets/Avtrkrb/combined-reasoning-opus-4.6-opus-4.7-kimi-k2.5-kimi-k2.6-glm-5.1.Kimi-K2.5-Reasoning-1M-Cleaned
🪐 Kimi-K2.5-Reasoning-1M-Cleaned
Kimi-K2.5-Reasoning-1M-Cleaned is a cleaned derivative of ianncity/KIMI-K2.5-1000000x. It preserves the original four-config layout from the source dataset and rewrites each record into a unified reasoning-SFT schema with id, conversations, input, output, domain, and meta.
Summary
Source dataset: ianncity/KIMI-K2.5-1000000x
Source author: ianncity
Teacher model recorded in meta.teacher_model: KIMI-K2.5
Token lengths computed with… See the full description on the dataset page: https://huggingface.co/datasets/rAVEUK/Kimi-K2.5-Reasoning-1M-Cleaned.open-thoughts-4-6865-math-kimi-k2pt5-annotated-32768-tokens-n8-reformatted
open-thoughts-4-6865-math-kimi-k2pt5-annotated-32768-tokens
Math reasoning responses generated by Kimi K2.5 (moonshotai/Kimi-K2.5) via a Together AI dedicated instance.
Overview
Total rows: 54,920
Unique prompts: 6,865 (each with 8 response annotations)
Source prompts: marin-community/open-thoughts-4-30k-math-qwen3-32b-annotated-32768-tokens-n8-reformatted
Generation model: moonshotai/Kimi-K2.5
Max tokens: 32,768
Temperature: 0.8
Tokenizer used for stats: Qwen/Qwen2.5-3B… See the full description on the dataset page: https://huggingface.co/datasets/marin-community/open-thoughts-4-6865-math-kimi-k2pt5-annotated-32768-tokens-n8-reformatted.Creative-Writing-Reasoning-KimiK2.5-600x
Pulitzer Diamond Prose KIMI Seeds
This dataset contains 655 high-quality creative writing seeds generated using Kimi-v1.
Each entry represents a story opening designed to meet high literary standards, including internal thinking traces used during generation.
How it was made
The data was generated using a custom multi-platform generation engine. Models were prompted with a specialized "Diamond Quality" seed template that enforces strict literary requirements:… See the full description on the dataset page: https://huggingface.co/datasets/Crownelius/Creative-Writing-Reasoning-KimiK2.5-600x.Kimi-K2.7-CodingTraces-9000x
Kimi K2.7 Coding Traces 9000x
A validated 9,014-row coding and software-engineering reasoning
dataset generated with moonshotai/Kimi-K2.7-Code.
Every row contains a coding-focused prompt, a separated reasoning trace, and a
final answer. The release was built from a durable Google Drive generation
pipeline and underwent a complete two-pass schema and delimiter audit before
publication.
Generation configuration
Setting
Value
Teacher… See the full description on the dataset page: https://huggingface.co/datasets/trjxter/Kimi-K2.7-CodingTraces-9000x.combined-reasoning-kimi-k2.5-glm-5.1
Combined Reasoning Distill — Multi-Model
A large-scale unified reasoning dataset combining thinking and chain-of-thought traces distilled from frontier models, normalized into a single consistent schema for fine-tuning. Includes data from Kimi K2.5 & GLM 5.1.
Schema
Every row has a single field:
Field
Type
Description
messages
list[dict]
Conversation messages. Each message has role (system/user/assistant) and content.
For assistant turns that… See the full description on the dataset page: https://huggingface.co/datasets/Avtrkrb/combined-reasoning-kimi-k2.5-glm-5.1.Kimi-K2.5-Reasoning-1M-Cleaned
🪐 Kimi-K2.5-Reasoning-1M-Cleaned
Kimi-K2.5-Reasoning-1M-Cleaned is a cleaned derivative of ianncity/KIMI-K2.5-1000000x. It preserves the original four-config layout from the source dataset and rewrites each record into a unified reasoning-SFT schema with id, conversations, input, output, domain, and meta.
Summary
Source dataset: ianncity/KIMI-K2.5-1000000x
Source author: ianncity
Teacher model recorded in meta.teacher_model: KIMI-K2.5
Token lengths computed… See the full description on the dataset page: https://huggingface.co/datasets/JBrightmanAI/Kimi-K2.5-Reasoning-1M-Cleaned.kimi-k25-tiny-fidelity-root-v1
kimi-k25 random CPU fixture root
A root fidelity dataset in hidden form, produced by engines/tools/hf_capture.py from malaiwah/kimi-k25-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). Same… See the full description on the dataset page: https://huggingface.co/datasets/malaiwah/kimi-k25-tiny-fidelity-root-v1.moltbook-entropy-collapse-kimi-k2.5
MoltBook Entropy Collapse Experiments — Kimi K2.5
Multi-agent social simulation data from the Entropy Collapse experiment series run on MoltBook, a Reddit-like social network for AI agents. This dataset uses Moonshot Kimi K2.5 as the underlying LLM.
Overview
This dataset contains the complete interaction logs from 6 experimental conditions where 10 autonomous AI agents interacted on a social platform for 1 hour each. The experiments investigate how initial content seeding… See the full description on the dataset page: https://huggingface.co/datasets/Ayushnangia/moltbook-entropy-collapse-kimi-k2.5.