One row per AI lab; part of a comparable battery across…: When will Safe Superintelligence (SSI) first have a model…

Part of: One row per AI lab; part of a comparable battery across new AGI-style labs with multibillion-dollar war chests, gauging whether any could rival OpenAI or Anthropic. Question per row: when will the named lab first have a frontier model? Resolution: the earliest date at which either (a) a model from the lab appears in the top 5 of the Artificial Analysis Intelligence Index (or its successor flagship overall-quality leaderboard) at any snapshot, or (b) for unreleased or unlisted systems, credible reporting plus at least one independent expert assessment establishes that the lab has trained a model broadly regarded as frontier-class, comparable to the contemporaneous top models from OpenAI, Anthropic, or Google (the way Anthropic's unreleased Mythos-class model inside Project Glasswing was publicly established before launch). If the lab is acquired or wound down first, or the event seems unlikely for many years, express that through late percentiles. Anchors: the current index top-5 is occupied by Anthropic and OpenAI variants; frontier generations currently require training runs costing billions of dollars and multi-hundred-megawatt compute. — view all 6 rows

When will Safe Superintelligence (SSI) first have a model at the AI frontier, per the resolution criteria?✓#1

frontier_dateJan 15, 2029

10% by
Mar '27
25% by
Jan '28
median
Jan '29
75% by
Mar '31
90% by
Jun '34

rationale:

Current Status & Resolution Mechanics Safe Superintelligence (SSI) remains a stealth research lab with no released model or product. The Artificial Analysis Intelligence Index top-5 is currently occupied entirely by Anthropic and OpenAI variants requiring mid-to-high-50s scores artificialanalysis.ai. Given SSI’s explicit "straight-shot, no products before superintelligence" doctrine 2 sources, a public leaderboard entry via an externally testable API (prong a) is highly unlikely. Instead, resolution will almost certainly occur via prong (b): credible reporting combined with an independent expert assessment that an internal SSI model is contemporaneous-frontier-class.

The Resource Picture & NVIDIA Scale-Up SSI is exceptionally well-funded for a new lab, having raised $1B initially and another $2B at a $32B valuation in early 2025 2 sources. While early experimentation reportedly relied heavily on Google Cloud TPUs 2 sources, a major July 2026 strategic partnership with NVIDIA marked a significant structural shift. The deal, which reportedly included a ~$5B equity investment, granted SSI access to the next-generation Vera Rubin platform and an "order of magnitude" increase in compute nvidianews.nvidia.com. However, NVIDIA’s statement that SSI has "research that is worthy of scaling up" is an endorsement of a research direction from a commercial partner and investor, not an independent expert assessment of an existing frontier model.

Short-Term Rumors vs. The Frontier Bar Recent rumors, such as a prominent investor's podcast claim that SSI plans to release a model in August 2026 2 sources, offer very weak evidence for a near-term resolution. This secondhand claim is uncorroborated, contradicts SSI’s stated non-product strategy fourweekmba.com, and focuses on sample-efficient learning—an axis that may not translate to a top-5 score on an agentic/coding-weighted index. Furthermore, a first release from a lab that has been comparatively compute-poor is unlikely to immediately rival Claude Opus 5 or GPT-5.6 Sol. The closest analogue, Thinking Machines' Inkling, debuted roughly 20 points off the frontier artificialanalysis.ai, illustrating the extreme difficulty of cracking the current top tier. This dynamic keeps early percentile dates relatively conservative.

Catching a Moving Target The primary challenge for SSI is that frontier capabilities are a rapidly receding target. While the new Vera Rubin hardware provides a massive compute upgrade, SSI is competing against incumbents preparing multi-billion dollar training runs and multi-gigawatt data centers 3 sources. SSI must not only execute a massive, multi-year training effort—navigating the friction of transitioning from Google TPUs to NVIDIA GPUs—but also catch up to competitors who benefit from enormous deployment feedback loops. This structural reality pushes the median expectation to early 2029, allowing time for SSI to complete at least one post-scale-up mega-run and for subsequent leaks or vetting to satisfy the independent assessment criteria.

Structural Risks and the Long Tail The right tail extends well into the 2030s due to several converging risks. SSI’s extreme secrecy and deliberate lack of public benchmarking could delay independent establishment for years, even if they achieve internal success. Furthermore, there is a substantial probability that SSI's non-scaling research bets fail to hit the contemporaneous frontier, or that the lab is eventually acquired or absorbed by a larger tech player—a scenario the forecasting criteria dictate expressing via very late percentiles.