Show HN: Gemini 4 Argon plays StarCraft

A benchmark called StarSkirmish Bench has been updated with results from the flagship models of OpenAI, Anthropic, Google, and Muse, showing how well they play StarCraft: Brood War by writing code. The models are Gemini 4 Argon from OpenAI, Llama from Anthropic, PaLM 2 from Google, and Muse 2 from Muse. The benchmark measures how well the models can use code to play the game, and the results show that each model has its strengths and weaknesses in different areas of the game.