Dataset Library

Reasoning traces for distilling frontier models

Curated datasets built by querying Claude, GPT, Gemini and other frontier models with diverse coding, math, and reasoning prompts. Designed for training small open models that still think clearly.

What's included

Each dataset includes detailed reasoning traces, carefully filtered conversations, and metadata ready for fine-tuning. Listings are synced hourly from Hugging Face.

Source:
35 datasets

claude-4.5-opus-high-reasoning-250x

Distilled from Claude Opus 4.5

SIZE<1KJSONTEXT
8.4K downloads160 likes

gemini-3-pro-preview-high-reasoning-1000x

Distilled from Gemini 3 Pro

SIZE1K–10KJSONTEXT
1.6K downloads56 likes

gpt-5.1-high-reasoning-1000x

Distilled from GPT-5.1

SIZE1K–10KJSONTEXT
618 downloads14 likes

gpt-5.2-high-reasoning-250x

SIZE<1KJSONTEXT
451 downloads4 likes

claude-haiku-4.5-high-reasoning-1700x

SIZE1K–10KJSONTEXT
404 downloads1 likes

gemini-3-flash-preview

SIZE10K–100KJSONTEXT
388 downloads7 likes

claude-sonnet-4.5-high-reasoning-250x

Distilled from Claude Sonnet 4.5

SIZE<1KJSONTEXT
381 downloads28 likes

glm-4.7-2000x

SIZE1K–10KJSONTEXT
329 downloads67 likes

deepseek-v3.2-speciale-openr1-math-3k

Distilled from DeepSeek v3.2 Speciale

SIZE1K–10KJSONTEXT
318 downloads4 likes

deepseek-v3.2-speciale-1000x

Distilled from DeepSeek v3.2 Speciale

SIZE<1KJSONTEXT
305 downloads5 likes

gpt-5.1-codex-max-1000x

Distilled from GPT-5.1

SIZE1K–10KJSONTEXT
260 downloads2 likes

gpt-5-codex-1000x

Distilled from GPT-5 Codex

SIZE<1KJSONTEXT
225 downloads1 likes

deepseek-v3.2-speciale-OpenCodeReasoning-3k

Distilled from DeepSeek v3.2 Speciale

SIZE1K–10KJSONTEXT
196 downloads6 likes

gpt-5-codex-250x

Distilled from GPT-5 Codex

SIZE<1KJSONTEXT
184 downloads8 likes

gemini-3-pro-preview-high-reasoning-250x

Distilled from Gemini 3 Pro

SIZE<1KJSONTEXT
181 downloads6 likes

MiniMax-M2.1-8800x

SIZE1K–10KJSONTEXT
159 downloads6 likes

MiMo-V2-Flash-2300x

SIZE1K–10KJSONTEXT
150 downloads2 likes

gemini-3-flash-preview-1000x

SIZE1K–10KJSONTEXT
144 downloads3 likes

minimax-m2.1-1000x

SIZE<1KJSONTEXT
135 downloads1 likes

grok-code-fast-1-1000x

Distilled from Grok

SIZE1K–10KJSONTEXT
119 downloads4 likes

glm-4.6-250x

Distilled from GLM 4.6

SIZE<1KJSONTEXT
115 downloads4 likes

kimi-k2-thinking-1000x

Distilled from Kimi K2

SIZE<1KJSONTEXT
101 downloads5 likes

gemini-2.5-flash-11000x

Distilled from Gemini 2.5 Flash

SIZE10K–100KTEXT
89 downloads4 likes

claude-haiku-4.5-1700x

SIZE1K–10KJSONTEXT
81 downloads0 likes

kimi-k2-thinking-250x

Distilled from Kimi K2

SIZE<1KJSONTEXT
70 downloads3 likes

polaris-alpha-1000x

SIZE1K–10KJSONTEXT
58 downloads3 likes

glm-4.7-350x

SIZE<1KJSONTEXT
57 downloads2 likes

gemini-3-flash-preview-standalone-html-1k

SIZE1K–10KJSONTEXT
56 downloads0 likes

brainstorm-v3.1-grok-4-fast-200x

Distilled from Grok

SIZE<1KJSONTEXT
55 downloads0 likes

sherlock-thinking-alpha-11000x

SIZE10K–100KJSONTEXT
49 downloads0 likes

open-moderator-v1

SIZE10K–100KJSONTEXT
42 downloads0 likes

gemini-2.5-flash-lite-2509-preview-1000x

Distilled from Gemini 2.5 Flash

SIZE<1KJSONTEXT
34 downloads1 likes

sherlock-think-alpha-1000x

SIZE1K–10KJSONTEXT
34 downloads1 likes

sherlock-dash-alpha-1000x

SIZE1K–10KJSONTEXT
31 downloads0 likes

convo-v1

26 downloads1 likes