finding
active
finding:gemini-3-flash-wins-67-9-of-its-28-mixed-format-games-against-code-agentsGemini 3 Flash wins 67.9% of its 28 mixed-format games against code agents
In the 172-game exp2 slice, G3-F has the highest LLM win rate against deterministic baselines.
Source paper
extracted_from(2026) · Robert Müller · Clemens Müller
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- G3-F achieved a win rate of 72.9% in the combined-comp1 98-game slice.
- Robust performance against algorithmic baselines.
- G3-F's median score across games was 5,250 points.
- G3-F completed on average 3.96 quartets per game.
- The Bayesian skill posterior mean for G3-F is 30.1 with 3σ confidence interval ±3.3.
- Poor performance against code agents.
- G3-F escalates bidding by nearly 10× from early to late game, adapting to phase.
- A trace shows G3-F turning a nominally wasteful overpay into a net score gain due to the multiplicative formula.