# DeepWism/DeepWism-R2

DeepWism R2 is a next-generation AGI system built on the T3CEDS framework (Thin-Thick-Thin Crowd Entropy Dynamics System), which redefines intelligence as a process of entropy reduction rather than attention modeling.

Repository: https://github.com/DeepWism/DeepWism-R2
Canonical: https://ross.abutalabs.com/products/deepwism-r2
License Family: other
Last push: 2025-06-27T03:20:06+00:00

## Health v2 (maintenance only)
Score: 31/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 28, release rhythm 35, longevity 31
- inputs: {"age_days": 436, "days_push": 432, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1017, forks 154 (observed 2026-08-28T04:03:14.771908+00:00)

## What it is
DeepWism R2 is an AI 'Research&Report' system presented as a next-generation AGI agents framework built on a Thin-Thick-Thin Crowd Entropy Dynamics System (T3CEDS), claiming entropy reduction rather than attention modeling as its core mechanism. It offers a chat interface for complex reasoning, deep retrieval, and automated research report generation, with self-reported benchmark results on Humanity's Last Exam and xbench.

## Use cases
- generate a deep research report on a complex topic automatically
- AI agent that performs multi-step web research and answers hard science questions
- chat assistant for deep search and reasoning across multiple domains
- explore an entropy-based alternative to attention-based AI architectures
- agent for multi-domain question answering across science, retrieval, and logic
- research agent benchmarked against OpenAI DeepResearch

## When to choose
- You want an autonomous research-and-report agent that combines deep retrieval with structured reasoning and produces reports
- You want to experiment with an alternative architecture built on entropy reduction and crowd-intelligence mechanisms
- You need a chat interface for complex science QA and deep search tasks

## When to avoid
- You need production software with a clear license - the repository has no license and an unknown primary language
- You require independently verified performance - the claimed HLE and xbench scores are self-published with limited methodological detail
- You need documented, inspectable implementation code - the README offers mostly high-level marketing claims about the T3CEDS architecture
- You prefer standard, well-supported LLM and agent tooling with active community and vendor support

## Facets
- artifact type: application
- maturity: experimental
- function: agent-framework, rag, chatbot, nlp
- domain: artificial-intelligence, chatbots, large-language-models
- platform: cloud
- tags: deep-research, research-agent, agi, entropy-reduction, crowd-intelligence, t3ceds, research-reports, deep-search, self-reported-benchmarks, ai-agents, retrieval-augmented-generation, natural-language-processing, web-server

## Member repositories
- DeepWism/DeepWism-R2 (main) score 31

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:14.771908+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:09.413522+00:00, confidence not recorded.
  - readme: https://github.com/DeepWism/DeepWism-R2 (fetched 2026-08-28T04:03:14.771908+00:00, sha 31c7382a8445)
- Data as of 2026-08-30T08:39:29.467469+00:00.