Fluent operation across text, image, audio and video.
Methodology v1.2 · updated Sep 2026
How close are we to artificial general intelligence?
The Index is a weighted composite of twelve capability dimensions, scored against an explicit operational threshold. It is an analytical estimate under a published framework — not a measurement, because no measurement of this exists.
63.4
Composite score out of 100
+2.7 since Aug 2026
Trajectory
slowing
Forecast confidence
medium
Central estimate
2032
Low-confidence weight
36%
Progression
The Index over time
Reconstructed under the current methodology version. Because the framework changed, historical values are back-cast rather than contemporaneous — a limitation we would rather state than hide.
| Date | Index |
|---|---|
| Jan 2020 | 18.5 |
| Jul 2020 | 21.0 |
| Jan 2021 | 24.2 |
| Jul 2021 | 27.0 |
| Jan 2022 | 30.4 |
| Jul 2022 | 33.1 |
| Jan 2023 | 39.8 |
| Jul 2023 | 43.5 |
| Jan 2024 | 47.9 |
| Jul 2024 | 51.2 |
| Jan 2025 | 55.4 |
| Jul 2025 | 58.6 |
| Jan 2026 | 61.0 |
| Apr 2026 | 62.1 |
| Jul 2026 | 62.9 |
| Sep 2026 | 63.4 |
Component scores
Twelve dimensions
Sorted by score. The hairline on each bar marks last month's position. Every dimension links to its evidence, its counter-evidence and what would change it.
- Multimodality82+2Highest dimension
- Coding81+3
- Tool use76+4
- Reasoning74+4
- Memory67+3
- Generalization66+3
- Planning61+3
- Learning59+1
- Autonomy53+3
- Social intelligence52+2
- Scientific discovery48+3
- Embodiment41±0Lowest dimension
| Dimension | Score | Previous | Change | Confidence |
|---|---|---|---|---|
| Multimodality | 82 | 80 | +2 | high |
| Coding | 81 | 78 | +3 | high |
| Tool use | 76 | 72 | +4 | high |
| Reasoning | 74 | 70 | +4 | medium |
| Memory | 67 | 64 | +3 | medium |
| Generalization | 66 | 63 | +3 | low |
| Planning | 61 | 58 | +3 | medium |
| Learning | 59 | 58 | +1 | low |
| Autonomy | 53 | 50 | +3 | medium |
| Social intelligence | 52 | 50 | +2 | low |
| Scientific discovery | 48 | 45 | +3 | low |
| Embodiment | 41 | 41 | 0 | medium |
Coding
81Writing, reading and repairing software.
Tool use
76Operating external systems to extend capability.
Multi-step inference that holds together over long chains.
Memory
67Retention and retrieval across long interactions.
Transfer of competence to genuinely unfamiliar problems.
Planning
61Goal decomposition and recovery when a plan fails.
Learning
59Acquiring new capability after training ends.
Autonomy
53Useful operation without a human in the loop.
Modelling other minds and coordinating with them.
Generating findings that survive independent replication.
Acting competently in unstructured physical space.
Forecast
When could AGI arrive?
Scenario weights, not a prediction of a dated event. A different operational threshold moves every number on this page.
| Threshold | Scenario | Probability |
|---|---|---|
| by 2027 | Discontinuous | 5% |
| by 2030 | Accelerated | 18% |
| by 2035 | Central | 42% |
| by 2040 | Extended | 25% |
| 2050 or later | Conservative | 10% |
Accelerated
2028–2030 · 23%Reliability improves fast enough that autonomy crosses the deployment threshold, and systems begin contributing materially to their own improvement.
Requires
Central
2031–2035 · 42%Current trajectories hold. Digital capability continues compounding while embodiment and durable learning are solved more slowly, by ordinary research rather than by breakthrough.
Requires
Conservative
2040 and beyond · 35%One or more of the low-scoring dimensions turns out to require a genuine conceptual advance rather than scale. Learning and embodiment are the leading candidates.
Requires
Retrospective
Milestones and their Index impact
Scored retrospectively under the current framework. These are not contemporaneous measurements — nobody was running this Index in 2017.
- May 2026Closed-loop autonomous experimentationSAMPLE ENTRY. Propose, run, measure, revise — closed without human intervention in narrow materials domains. The question of who frames the research question remains open.Multiple institutionsscience+2.2
- Nov 2025Cross-embodiment manipulation transferSAMPLE ENTRY. Manipulation policies transferring across robot morphologies without retraining — the genuine advance in an otherwise static dimension.Multiple laboratoriesembodiment+1.9
- Mar 2025Repository-scale autonomous software engineeringSAMPLE ENTRY. Systems operating across large unfamiliar codebases rather than isolated functions — the shift that made the coding dimension economically visible.Multiple laboratoriesagents+3.8
- Sep 2024Inference-time deliberationSpending more computation at inference on harder problems produced gains that scaling alone had stopped delivering, and opened a second axis of improvement.Multiple laboratoriesreasoning+4.7
- Sep 2023Natively multimodal frontier modelsText, image and audio brought into shared representations rather than routed between specialist components — the change that took multimodality from a capability to an assumption.Multiple laboratoriesscaling+3.4
- Nov 2022Conversational deployment at consumer scaleMoved frontier capability from research artefact to daily instrument for hundreds of millions of people. The Index treats scaled deployment as reliability evidence that no benchmark provides.OpenAIscaling+5.1
- Nov 2020Protein structure prediction at experimental accuracyThe first case of a machine learning system resolving a long-standing open problem in the natural sciences, and still the strongest single piece of evidence in the scientific-discovery dimension.Google DeepMindscience+2.9
- May 2020Large-scale few-shot language modellingDemonstrated that capability could emerge from scale without task-specific training, reframing the research agenda around scaling rather than architecture.OpenAIscaling+3.6
- Jun 2017The transformer architecture is publishedAttention replaced recurrence, making training parallelisable and turning scale into a lever the field could actually pull. Nearly every system the Index tracks descends from this paper.Googlefoundations+4.2