Google reports a 96.2% F1 score for Gemini 4 Carbon on GraphWalks breadth-first-search problems with 256K to 1M tokens of context, which it ran itself.
Tasks: The 200 GraphWalks problems with contexts between 256K and 1M tokens; breadth-first search; F1 score.
Setting: Google says its results use the Gemini API at the highest thinking setting unless noted; no numeric budget is stated.
Limits: Provider-reported: Google publishes the result for its own model; it has not been independently reproduced here. Google says the scores are as of October 2026 for the model it calls Gemini 4 Carbon; equivalence to the build now served on every surface is not established. Results under the same benchmark name can be shown together as reported scores, but unmatched settings and missing methods rule out a controlled ranking. The test is filed with its length band and task type in the name so that other GraphWalks subsets are never pooled with it.
Google DeepMind · Google DeepMind · Google DeepMind
