Anthropic's New AI Model Tops 50% on All Benchmarks in METR Analysis
Key Takeaways
- •Model 4, Anthropic's latest internal AI system as identified by METR, scored above 50% on every benchmark tested, a first in the evaluation organization's dataset.
- •The model finished training last month and has become the highest-scoring model in METR's analysis, indicating swift progress in Anthropic's AI capabilities.
- •Current market pricing aligns with heightened expectations that Anthropic will hold the best AI model by the end of November 2026, amid competition from Google, Meta, and OpenAI.
- •Anthropic has not connected Model 4 to any commercial product but announced Claude Opus 5.5, Sonnet 5.5, and Haiku 5.5 on the same day, without confirming whether Model 4 is tied to that lineup.
- •Observers expect further Anthropic announcements, rival responses, and new independent evaluations ahead of December 2026 to indicate which company may lead in AI model performance.
Anthropic's latest internal AI model, referred to as Model 4 by AI evaluation organization METR, has scored above 50% on every benchmark it was tested on, according to a recent METR analysis. The organization evaluates frontier AI systems across a range of capability benchmarks. The result marks the first time in METR's dataset that a model has cleared the 50% threshold on all benchmarks tested.
The model completed its training last month and has quickly become the highest-scoring model in METR's analysis, a development that points to significant advancements in Anthropic's AI capabilities and reflects rapid progress compared with previous periods. Because METR's benchmark suite spans a range of capability areas, the result reflects performance across the organization's full set of tests rather than strength on any single measure.
The milestone carries implications for Anthropic's standing in the increasingly competitive AI landscape. Market participants are currently adjusting their expectations over which company will hold the best AI model by the end of November 2026, and current market pricing appears consistent with increased expectations for Anthropic's AI leadership over that period. Anthropic's gains are unfolding amid a broader technological race among the leading AI firms — including Google, Meta, and OpenAI — each of which is vying for dominance in the field.
Anthropic has yet to publicly associate Model 4 with any of its commercial products. However, the company announced its Claude Opus 5.5, Sonnet 5.5, and Haiku 5.5 models on the same day, a move that indicates strategic alignment between its public-facing product releases and its internal model development efforts. Anthropic has not disclosed whether Model 4 is connected to the newly announced Claude 5.5 lineup.
What to Watch
According to the analysis, market participants are likely to monitor for further announcements from Anthropic regarding the public release or application of Model 4. Competitive responses from other major AI companies, such as Google and Meta, could also influence market dynamics in the period ahead. New benchmarks or independent evaluations of AI models released in the lead-up to December 2026 are expected to serve as key indicators of which company may ultimately claim the title of having the best AI model.
