Not even Pokémon is safe from AI benchmarking controversy.

Last week, a post on X went viral, claiming that Google’s latest Gemini model surpassed Anthropic’s flagship Claude model in the original Pokémon video game trilogy. Reportedly, Gemini had reached Lavendar Town in a developer’s Twitch stream; Claude was stuck at Mount Moon as of late February.

Gemini is literally ahead of Claude atm in pokemon after reaching Lavender Town

119 live views only btw, incredibly underrated stream pic.twitter.com/8AvSovAI4x

— Jush (@Jush21e8) April 10, 2025

But what the post failed to mention is that Gemini had an advantage.

As users on Reddit pointed out, the developer who maintains the Gemini stream built a custom minimap that helps the model identify “tiles” in the game like cuttable trees. This reduces the need for Gemini to analyze screenshots before it makes gameplay decisions.

Now, Pokémon is a semi-serious AI benchmark at best — few would argue it’s a very informative test of a model’s capabilities. But it is an instructive example of how different implementations of a benchmark can influence the results.

For example, Anthropic reported two scores for its recent Anthropic 3.7 Sonnet model on the benchmark SWE-bench Verified, which is designed to evaluate a model’s coding abilities. Claude 3.7 Sonnet achieved 62.3% accuracy on SWE-bench Verified, but 70.3% with a “custom scaffold” that Anthropic developed.

More recently, Meta fine-tuned a version of one of its newer models, Llama 4 Maverick, to perform well on a particular benchmark, LM Arena. The vanilla version of the model scores significantly worse on the same evaluation.

Given that AI benchmarks — Pokémon included — are imperfect measures to begin with, custom and non-standard implementations threaten to muddy the waters even further. That is to say, it doesn’t seem likely that it’ll get any easier to compare models as they’re released.