How Close Are We to AGI?

Capability jumps are real — agents, multimodal models, and research systems keep raising the ceiling. AGI, however, is a contested definition. Expect continued frontier progress and louder AGI claims without a clean, consensus crossing date in the near term.

AGI timelines drive capital allocation, safety policy, chip demand, and public expectations. A clearer read on proximity helps separate durable signals from hype cycles.

• Frontier capability benchmarks keep rising

  New model generations improve on coding, science, and long-horizon tasks.

  Signal · strong

• Definitions remain contested

  Labs, academics, and policymakers still disagree on what counts as AGI.

  Signal · strong

• Autonomy demos outpace reliable deployment

  Impressive demos coexist with production constraints around verification and cost.

  Signal · moderate

• Safety and governance lag claims

  Policy and evaluation frameworks trail the pace of public AGI rhetoric.

  Signal · moderate

• Scenario A

  Rapid threshold crossing

  A lab demonstrates broadly general, economically transformative systems and consensus forms quickly.

• Scenario B

  Incremental generality

  Capabilities compound for years without a single AGI declaration; products feel more general while the label stays disputed.

• Scenario C

  Plateau narrative returns

  Scaling or data constraints slow progress; AGI discourse cools relative to applied AI.

• OpenAI
• Google DeepMind
• Anthropic
• Meta
• xAI

• company

  Public model release cadence and eval reports

• research

  Academic surveys of AGI definitions and timelines

• policy

  Government AI strategy and safety documents

1. Sep 22, 202644%
2. Sep 15, 202642%
3. Sep 8, 202640%
4. Sep 1, 202638%

• Will AI Agents Replace Traditional Software?
• Will Open-Weight AI Models Catch Up?
• Is China Catching Up with the US in AI?