SSI has released no models, papers, or products, and a rumored August 2026 launch never materialized https://www.startuphub.ai/ai-news/ai-figures/2026/figure-ilya-sutskever-ssi-first-model-2026-08-14 [https://www.startuphub.ai/ai-news/ai-figures/2026/figure-ilya-sutskever-ssi-first-model-2026-08-14](https://www.startuphub.ai/ai-news/ai-figures/2026/figure-ilya-sutskever-ssi-first-model-2026-08-1
Evaluating SSI's timeline in the context of the severe structural headwinds facing all challenger labs attempting to match incumbent scale pushed the distribution slightly later by a few months.
Thinking Machines Lab (TML) is currently far from the AI frontier. As of October 2026, its best models, Inkling and Inkling-Small, score 25 and 26 respectively on the Artificial Analysis Intelligence Index v4.3.2. This severely trails the top five entries, which score between 54 and 58 https://artificialanalysis.ai/leaderboards/models, and significantly lags top open-weights models like Xiaomi's MiMo-V2.6-Pro at 46 [https://artificialanalysis.
Evaluating the immense structural hurdles that fast-following AI challenger labs face in catching up to the capabilities of well-resourced incumbents shifted the estimated timeline significantly later across all percentiles.
Discovery Loop is highly unlikely to have a frontier AI model for many years, and may never do so. Its stated mission is to automate the experimental loop of machine-learning research, rather than to build general-purpose chatbots Discovery Loop — Continuous Exploration Jeff Dean leaves Google after 27 years to build an AI company that ... [Jeff Dean
Modeling the immense structural hurdles facing new challenger labs and the high likelihood of acquisition or narrow specialization shifted the projected arrival of a general frontier model back by several years and increased the probability it never happens.
Reflection AI's path to the frontier is heavily gated by a severe performance deficit and looming acquisition risks. Although the lab possesses the financial and compute war chest needed to scale, its inaugural open-weight model is deeply behind both proprietary flagships and top-tier open models. Consequently, the estimates reflect a prolonged timeline: the 10th and 25th percentiles fall in June 2028 and September 2029, while the median is pushed to October 2031 and the 75th percentile to Janua
Accounting for the severe difficulty of matching incumbent general models compared to merely leading the open-weights category, and recognizing that an acquisition could permanently end the lab's independent path to the frontier, shifted the furthest tail of this timeline to "never".
Ineffable Intelligence, founded in late 2025 by DeepMind's former reinforcement-learning lead David Silver, has substantial backing but no models, papers, or products to date Ineffable Intelligence David Silver's Ineffable Names Six Cofounders . The lab raised a $1.1B seed at a $5.1B valuation in April 2026 [DeepMind's David Silver just raised $1.1B to build an AI ...](https:
Evaluating the structural hurdles facing non-standard challenger labs compared to the immense resources of incumbents pushes the median arrival date slightly later into 2036 and strictly caps the upper percentiles at 'never' to reflect a high, persistent risk of early acquisition or failure to reach broad general capabilities.
AMI Labs is unlikely to produce a broadly defined frontier model in the near term, if it ever does as an independent entity. As of October 2026, the lab has released no public model and is entirely absent from the Artificial Analysis Intelligence Index Advanced Machine Intelligence Labs https://artificialanalysis.ai/leaderboards/models. Backed by a $1.03B seed round, the lab's
Comparing AMI Labs' non-LLM trajectory against the broader development hurdles for alternative and challenger AI labs pushed the estimated timeline for reaching broad frontier capabilities slightly later.
Ask a followup
Sign in to run · $20 free credit, no card · every claim cited