{"event":{"id":"evt-opt-175b-20220503","dedupe_key":"controlled_frontier_model_release:2022-05-03:meta-releases-opt-175b-for-controlled-research-access","event_type":"controlled_frontier_model_release","title":"Meta releases OPT-175B for controlled research access","summary":"Meta released code, smaller checkpoints, training logs and controlled noncommercial research access to OPT-175B, bringing 175B-scale model inspection closer to researchers outside a few frontier labs.","occurred_at":"2022-05-03T00:00:00.000Z","published_at":"2022-05-03T00:00:00.000Z","observed_at":"2026-10-01T18:14:00.000Z","reconstructed_at":"2026-10-01T18:14:00.000Z","ingest_type":"live_world_scan","locations":null,"status":"verified","confidence":0.99,"metadata":{"actors":["Meta AI"],"what_happened_at_time":"A 175B-scale language model became available to approved researchers under a noncommercial license, alongside extensive training documentation and open software.","evidence_at_time":"The May 2 preprint and May 3 Meta release documented access terms, compute requirements, training logs and stated carbon-footprint comparisons.","affected_layers":["layer-models","layer-open-source","layer-research","layer-software","layer-compute","layer-governance","layer-energy"],"change_kind":"access_bottleneck_reduction","q1_continuity":"Q1 GPT-NeoX-20B made open/public large-model access material. OPT pushed research access much closer to then-frontier parameter scale but remained gated and licensed.","retrospective_inference":"The open-model bottleneck shifted from simple absence of large weights toward governance, licensing, serving cost and compute access.","uncertainty":"Calling OPT-175B fully open source would overstate the release: the largest checkpoint used controlled, noncommercial access."},"created_at":"2026-10-01T19:22:17.747Z","updated_at":"2026-10-01T19:22:17.747Z"},"artifacts":[{"id":"art-opt-175b-20220503","canonical_url":"https://ai.meta.com/blog/democratizing-access-to-large-scale-language-models-with-opt-175b/","artifact_type":"controlled_model_release","title":"OPT-175B research release","summary":"Meta released OPT model code, smaller checkpoints and controlled noncommercial access to OPT-175B, alongside training logs and infrastructure notes intended to broaden research access to frontier-scale language models.","creator_entities":["Meta AI"],"released_at":"2022-05-03T00:00:00.000Z","content_hash":null,"metadata":{"historical_scope":"2022-Q2","openness_note":"175B access was controlled and noncommercial; this was not unrestricted open-source release."},"created_at":"2026-10-01T19:22:17.740Z","updated_at":"2026-10-01T19:22:17.740Z","relation":"documented_by"}],"signals":[{"id":"sig-q2-energy-carbon-metric","statement":"Energy and carbon intensity became explicit metrics in AI infrastructure and large-model training announcements, but Q2 evidence does not establish electricity supply as a binding AI compute bottleneck.","signal_type":"infrastructure_metric_visibility","direction":"increasing","confidence":0.83,"epistemic_status":"supported_inference","mapping_mode":"retrospective","reconstructed_at":"2026-10-01T18:14:00.000Z","created_at":"2026-10-01T19:22:17.756Z","updated_at":"2026-10-01T19:22:17.756Z"},{"id":"sig-q2-frontier-research-access","statement":"Research access to very large language models improved sharply, but the access bottleneck shifted toward licenses, approval, serving cost and the compute needed to reproduce or run them.","signal_type":"access_bottleneck_shift","direction":"improving_with_constraints","confidence":0.96,"epistemic_status":"supported_inference","mapping_mode":"retrospective","reconstructed_at":"2026-10-01T18:14:00.000Z","created_at":"2026-10-01T19:22:17.753Z","updated_at":"2026-10-01T19:22:17.753Z"},{"id":"sig-q2-governance-stack-coupling","statement":"Public and private governance moved closer to the technical stack through AI-risk measurement work, semiconductor/export-control coordination and controlled model deployment/release practices.","signal_type":"governance_maturation","direction":"increasing","confidence":0.95,"epistemic_status":"supported_inference","mapping_mode":"retrospective","reconstructed_at":"2026-10-01T18:14:00.000Z","created_at":"2026-10-01T19:22:17.754Z","updated_at":"2026-10-01T19:22:17.754Z"}]}