DeepMind's internal GCI index shows steady ~0.1/month Gemini gains, no structural AI speedup
HaydnBelfield · x · 2026-10-14
Part 3/9 of DeepMind's report thread covers Finding 1: no significant end-to-end acceleration of ML R&D yet.
- DeepMind built GCI v1, an internal general capability index (similar to Epoch's ECI), fitted across 123 models and 92 benchmarks.
- Frontier Gemini capability has improved at a remarkably steady 0.1 GCI points/month.
- A structural break test on April 1, 2026 shows no statistically significant speedup (0.091 vs 0.118 pts/month on public models; 0.097 vs 0.107 including internal ones).
- Gemini 4 Argon sits above trend, but this is traced to increased compute and organizational investment rather than AI-driven acceleration.
Related event: Google launches Gemini 4 Argon across apps, API and Search(277 posts)→
More from Research
- Vals AI Puts Frontier Models in Real-Time MuJoCo Robot Simulations — scaling01 · 2026-10-15
- AI Post-Processing Tool Extends Extreme Weather Warnings From 10 Days to 6 Weeks — LesterMackey · 2026-10-15
- DuplexJev Cuts Autoregressive Decoding for Voice Agents, Answering 80 Speech Decisions in 0.1s — adventists-ai · 2026-10-15
- LLM-to-Flow Distillation Adapts Scientific Generators to Natural-Language Discovery Goals — rbhar90 · 2026-10-15
- ETH's Uncertainty-Aware World Model Makes Offline RL Work on Real Robots — breadli428 · 2026-10-15
- 167,356 simulated shop-days: paper shows eval world design writes the model behavior it scores — liminal_bardo · 2026-10-15