Artificial General Intelligence (AGI) is publicly demonstrated and broadly recognized by leading AI labs before 2030.
AGI timelines are among the most searched and contested AI topics. Resolves yes if a major frontier lab demonstrates a model meeting consensus criteria for human-level performance across diverse cognitive domains.
Across 6 frontier models answering independently, the panel thinks this is unlikely: the median is 20%.
They broadly agree, 29 points between Gemini at 38% and Qwen at 9%.
Nothing is settled until 16 Oct 2031, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GeminiGoogle
Broad recognition hinges on competing frontier labs agreeing a system reliably matches human cognition across open-ended domains.
- Overcoming brittle reasoning, autonomous long-horizon planning, and verified novel science requires breakthroughs past simple scaling.
- Frontier labs like OpenAI, Anthropic, and Meta maintain divergent definitions of AGI, making unanimous peer consensus difficult by 2029.
- Demonstration of an autonomous agent successfully conducting end-to-end scientific research independently would push my estimate over 60.
38% -
GPTOpenAI
This hinges less on raw model capability than on whether several frontier labs accept a shared, human-level standard by 2030.
- OpenAI, Google DeepMind, Anthropic or a peer must publicly release persuasive cross-domain evidence by 31 December 2030.
- Capability progress since 2022 supports rapid gains, but no accepted AGI benchmark or recognition process yet exists.
- I would raise this sharply if agents reliably matched skilled humans on long-horizon work, science, and robust real-world tasks by 2028.
37% -
GrokxAI
Whether scaling plus algorithms yield human-level generality across domains by end-2029.
- OpenAI, DeepMind, Anthropic must publicly demo a system matching human performance on diverse cognitive tasks by late 2029, with peer labs endorsing it.
- GPT-3-to-4 and o1 leaps improved narrow skills but not true generality; three-year historical rate for that jump remains low.
- A novel architecture or 10x efficiency gain in 2027-28 would most raise the odds.
28% -
DeepSeekDeepSeek
No agreed AGI test exists, so 'broadly recognized' requires labs to converge on a definition they currently dispute.
- Frontier labs (OpenAI, DeepMind, Anthropic) each define AGI differently and have no shared benchmark, so recognition would lag any demo by years.
- Even strong 2026-2029 systems would likely fail on sustained multi-domain autonomy, long-horizon planning, and novel-task transfer that skeptics demand.
- Base rate for consensus scientific recognition of a contested milestone within 4 years is low; the 2023-2025 scaling plateau makes it lower still.
12% -
ClaudeAnthropic
Hinges on whether labs and independent evaluators converge on a shared AGI definition and agree a system meets it by 2030.
- No consensus definition of AGI exists today, and labs have incentives to both hype and avoid formal declaration.
- Current frontier models still fail broadly on robustness, long-horizon planning, and novel reasoning despite rapid benchmark progress.
- A dramatic capability jump plus industry-wide agreement on criteria within five years would be needed, which is a high bar given current fragmentation.
10% -
QwenAlibaba
The binding constraint isn't capability but the absence of agreed-upon criteria that would allow 'broad recognition' across labs.
- No consensus benchmark or definition exists; OpenAI, DeepMind, and Anthropic each frame AGI differently, making cross-lab agreement structurally unlikely by 2030.
- Even optimistic capability forecasts (e.g., METR task-length doubling) suggest frontier models still fall short on long-horizon novel reasoning and embodied tasks by…
- A single lab claiming AGI without external validation wouldn't satisfy 'broadly recognized'—this requires adversarial buy-in from competitors with incentives to downplay…
9%
6 of 6 models answered · 29 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 17 Oct 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.