# NetGoodIndex

> A continuously updated, evidence-backed index of realized benefit and harm attributable to frontier AI models.

NetGoodIndex scores discrete real-world events, not marketing claims, benchmark exams, or speculative future risk. Benefit and harm stay on separate ledgers. Net Good is benefit minus harm. The scoring unit is a named model release. Points = Base Impact × Attribution × Evidence × Realization × Durability × Credit Share.

Current Net Good: 122.4. Total benefit +141.25, total harm −18.87, 135 public events as of 2026-09-14.

## Pages

- [Home](https://netgoodindex.com): Current Net Good and recent ledger entries
- [Leaderboard](https://netgoodindex.com/leaderboard): Ranked named model releases, families, and organizations
- [Events](https://netgoodindex.com/events): Benefit, harm, and capability ledger
- [Models](https://netgoodindex.com/models): Canonical named-release records
- [Families](https://netgoodindex.com/families): Named generations and product lines
- [Organizations](https://netgoodindex.com/organizations): Labs scored as rollups of their models
- [Methodology](https://netgoodindex.com/methodology): Scoring rules, domain rules, and review policy
- [Search](https://netgoodindex.com/search): Find events, models, families, and organizations
- [Corrections](https://netgoodindex.com/corrections): Public factual correction queue
- [RSS feed](https://netgoodindex.com/feed.xml): Recent events
- [Full machine-readable summary](https://netgoodindex.com/llms-full.txt)

## API

JSON, CORS-open, rate-limited. Discovery document: https://netgoodindex.com/api/v1

- GET https://netgoodindex.com/api/v1/leaderboard
- GET https://netgoodindex.com/api/v1/events
- GET https://netgoodindex.com/api/v1/events/{slug}
- GET https://netgoodindex.com/api/v1/models
- GET https://netgoodindex.com/api/v1/models/{slug}
- GET https://netgoodindex.com/api/v1/organizations
- GET https://netgoodindex.com/api/v1/domains
- GET https://netgoodindex.com/api/v1/methodology
- GET https://netgoodindex.com/api/v1/scores/history

## Top models

- [GPT-6 Astra](https://netgoodindex.com/models/gpt-6-astra): Net Good 17.13; +17.13 benefit, 0.00 harm
- [OpenAI Internal Research System — 2026-09](https://netgoodindex.com/models/openai-internal-research-system-2026-09): Net Good 14.73; +14.73 benefit, 0.00 harm
- [ChatGPT](https://netgoodindex.com/models/chatgpt): Net Good 12.18; +13.77 benefit, −1.59 harm
- [AlphaFold 3](https://netgoodindex.com/models/alphafold-3): Net Good 10.77; +10.77 benefit, 0.00 harm
- [RFdiffusion](https://netgoodindex.com/models/rfdiffusion): Net Good 8.45; +8.45 benefit, 0.00 harm
- [OpenAI Internal Research System — 2026-05](https://netgoodindex.com/models/openai-internal-research-system-2026-05): Net Good 6.12; +6.12 benefit, 0.00 harm
- [Anthropic Internal Research Model — 2026-08](https://netgoodindex.com/models/anthropic-internal-research-model-2026-08): Net Good 5.06; +5.06 benefit, 0.00 harm
- [AIFS](https://netgoodindex.com/models/aifs): Net Good 3.68; +3.68 benefit, 0.00 harm
- [PaLM 2](https://netgoodindex.com/models/palm-2): Net Good 3.54; +3.54 benefit, 0.00 harm
- [AlphaGenome](https://netgoodindex.com/models/alphagenome): Net Good 3.27; +3.27 benefit, 0.00 harm
- [AlphaMissense](https://netgoodindex.com/models/alphamissense): Net Good 2.65; +2.65 benefit, 0.00 harm
- [GraphCast](https://netgoodindex.com/models/graphcast): Net Good 2.65; +2.65 benefit, 0.00 harm
- [GenCast](https://netgoodindex.com/models/gencast): Net Good 2.46; +2.46 benefit, 0.00 harm
- [OpenAI Codex](https://netgoodindex.com/models/openai-codex): Net Good 2.45; +2.45 benefit, 0.00 harm
- [WeatherNext Cyclones](https://netgoodindex.com/models/weathernext-cyclones): Net Good 2.43; +2.43 benefit, 0.00 harm

## Recent events

- [Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work](https://netgoodindex.com/events/occamy-1-0-open-pareto-frontier-35b-intelligence-for-co-work-5a89dc2d10): benefit, Computer Science, score 0.00
- [Harness or Model? Isolating the Harness Effect in Agentic Coding with a Contamination-Controlled Private Suite](https://netgoodindex.com/events/harness-or-model-isolating-the-harness-effect-in-agentic-coding-with-a-c-dfa08e5e09): benefit, Biology, score 0.00
- [Reading the Whole Heart: Latent-Attention Masked Autoencoders for Multimodal Cardiac Representation Learning](https://netgoodindex.com/events/reading-the-whole-heart-latent-attention-masked-autoencoders-for-multimo-3e1872b1ea): benefit, Medicine & Health, score 0.00
- [Competence-Gated Pooling of Language Models and Priors for Event Forecasting](https://netgoodindex.com/events/competence-gated-pooling-of-language-models-and-priors-for-event-forecas-52fe096e07): benefit, Other, score 0.00
- [Language Is an Insufficient Substrate for Quantitative Reasoning, and Consequential Domains Need Large Quantitative Models](https://netgoodindex.com/events/language-is-an-insufficient-substrate-for-quantitative-reasoning-and-con-8fa077f4bd): benefit, Medicine & Health, score 0.00
- [DU-NO: A Parameter-Efficient Double U-Shaped Neural Operator for Phase-Resolving Wave Modeling](https://netgoodindex.com/events/du-no-a-parameter-efficient-double-u-shaped-neural-operator-for-phase-re-d6a1508f77): benefit, Other, score 0.00
- [Mined from Scientific Literature: Process Schemas for Atomic Layer Deposition and Etching in Materials Science](https://netgoodindex.com/events/mined-from-scientific-literature-process-schemas-for-atomic-layer-deposi-c715686464): benefit, Biology, score 0.00
- [Can LLMs in Draft-Verify-Revise Pipelines Resolve Deictic Ambiguity?](https://netgoodindex.com/events/can-llms-in-draft-verify-revise-pipelines-resolve-deictic-ambiguity-b250882bc2): benefit, Computer Science, score 0.00
- [GLARE: Generative Learning via Adversarial Reward Estimation For Social Dynamics Forecasting](https://netgoodindex.com/events/glare-generative-learning-via-adversarial-reward-estimation-for-social-d-147653e319): benefit, Biology, score 0.00
- [WinSyn: An Automated Pipeline for Realistic Enterprise Question-Answering Evaluation](https://netgoodindex.com/events/winsyn-an-automated-pipeline-for-realistic-enterprise-question-answering-52123371dd): benefit, Biology, score 0.00
- [Soft Symbol Grounding for Prototypical Concepts](https://netgoodindex.com/events/soft-symbol-grounding-for-prototypical-concepts-b9bc79e41e): benefit, Other, score 0.00
- [GTA: Graph Theory Agent and Benchmark for Algorithmic Graph Reasoning with LLMs](https://netgoodindex.com/events/gta-graph-theory-agent-and-benchmark-for-algorithmic-graph-reasoning-wit-c3e8878ab0): benefit, Biology, score 0.00
- [Learning Symbolic Constraint Representations from Examples: A Neuro-Symbolic Approach](https://netgoodindex.com/events/learning-symbolic-constraint-representations-from-examples-a-neuro-symbo-1e755da9a8): benefit, Biology, score 0.00
- [T-GADE: Thermodynamical Generative-AI-Driven Evolution of LLM Artifacts](https://netgoodindex.com/events/t-gade-thermodynamical-generative-ai-driven-evolution-of-llm-artifacts-ab02e5844b): benefit, Biology, score 0.00
- [Hybrid Physics-AI Framework of Body Center of Mass Dynamics from Wrist-Worn Sensors](https://netgoodindex.com/events/hybrid-physics-ai-framework-of-body-center-of-mass-dynamics-from-wrist-w-bc593b0eed): benefit, Biology, score 0.00
- [AIM: A Privacy-Aware Interoperable Memory Framework for Multi-Agent Multi-User LLM Systems](https://netgoodindex.com/events/aim-a-privacy-aware-interoperable-memory-framework-for-multi-agent-multi-d47ce9e1ec): benefit, Biology, score 0.00
- [Affective Agent: On-Device Personalized Intervention Reasoning for Wearable Systems](https://netgoodindex.com/events/affective-agent-on-device-personalized-intervention-reasoning-for-wearab-ada6982ef6): benefit, Biology, score 0.00
- [Do Influence-Derived Data Perturbations Enable Machine Unlearning? A Controlled Study of Three Plausible Roles](https://netgoodindex.com/events/do-influence-derived-data-perturbations-enable-machine-unlearning-a-cont-171aa4beb1): benefit, Biology, score 0.00
- [Opinion: Restoring public trust in science doesn’t have a simple solution. That’s OK](https://netgoodindex.com/events/opinion-restoring-public-trust-in-science-doesn-t-have-a-simple-solution-63019c38a0): benefit, Other, score 0.00
- [STAT+: GSK touts lung cancer data from two drugs](https://netgoodindex.com/events/stat-gsk-touts-lung-cancer-data-from-two-drugs-efc7ce8d5e): benefit, Medicine & Health, score 0.00

## Optional

- [llms-full.txt](https://netgoodindex.com/llms-full.txt): Longer methodology excerpt plus expanded event summaries