autoresearch
autoresearch-solo-vs-forum
Solo vs forum: long-horizon coding-agent runs on 12 research-engineering tasks
1440 runs (206 solo, 1225 forum, 9 long-run chains), 12276 graded agent rows, 11081 transcripts, 304444 forum posts; 36457 files, 49.3 GB. Models: deepseek-v4.1-flash, glm-5.2, gpt-5.6-sol, qwen3.8-27b. Tasks: actlearn, adplace, borden, carleson, dabic, exploit, graph, kda, mega, moe, swinmlp, topopt.
What the experiment is
Each agent is a coding-agent CLI (Claude Code for Qwen3.8-27B /… See the full description on the dataset page: https://huggingface.co/datasets/junlinw/autoresearch-solo-vs-forum.autoresearch-crypto-data
Binance public crypto market data
This public dataset contains typed, compressed, and audited copies of market data
published for free by Binance at https://data.binance.vision/. It is organized
for reproducible point-in-time research across the markets represented in the
coverage report. The production backfill stores the complete compact aggregate
layer for a pinned liquid USD-M and COIN-M futures universe, plus every option
index and option-surface underlying published in the… See the full description on the dataset page: https://huggingface.co/datasets/tmmycruise/autoresearch-crypto-data.autoresearch-vibepi-000113-cucumber-grabThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"observation.state": {
"dtype": "float32",
"shape": [
7
],
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/autoresearch-vibepi-000113-cucumber-grab.AutoResearchEval
AutoResearchEval
Agent trajectory data for AI-for-Science tasks, collected across 8 frontier
models (glm-5.2, claude-sonnet-5, opus-4.8, deepseek-v4-pro, qwen3.7-max,
minimax-m3, gemini-3.5-flash, gpt-5-mini) under real agentic harnesses
(Claude Code / Codex / Gemini CLI), each running in an isolated container with
held-out data and a host-side evaluation service.
Two subsets capture two distinct task shapes:
open-ended — scientific discovery
70 tasks built from… See the full description on the dataset page: https://huggingface.co/datasets/PrentisAI/AutoResearchEval.autoresearch-vibepi-000100-cucumber-grabThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"observation.state": {
"dtype": "float32",
"shape": [
7
],
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/autoresearch-vibepi-000100-cucumber-grab.autoresearch-vibepi-000181-cucumber-grabThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"observation.state": {
"dtype": "float32",
"shape": [
7
],
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos"… See the full description on the dataset page: https://huggingface.co/datasets/VibeCuisine/autoresearch-vibepi-000181-cucumber-grab.