Estebankiwi@estebankiwi

Takip et

Gemini 3.5 Flash SWE-Bench Pro Score

Evaluations reveal Gemini 3.5 Flash attaining 55.1% on SWE-Bench Pro while Claude Opus 4.7 reaches 64.3%. The gap proves substantial. Google created this Flash version that exceeds their prior Pro in tool usage along with agentic functions. Nevertheless in practical coding scenarios the model trails by nine points behind Opus 4.7. GPT 5.5 surpasses the Flash result with 58.6%. If this constitutes the key release for Google's return to prominence then shortcomings remain evident regarding coding ability. Anticipation builds for Gemini 3.5 Pro since that version will determine real capabilities.

3mo

Estebankiwi@estebankiwi

Takip et

Gemini 3.5 Flash SWE-Bench Pro Score

Evaluations reveal Gemini 3.5 Flash attaining 55.1% on SWE-Bench Pro while Claude Opus 4.7 reaches 64.3%. The gap proves substantial. Google created this Flash version that exceeds their prior Pro in tool usage along with agentic functions. Nevertheless in practical coding scenarios the model trails by nine points behind Opus 4.7. GPT 5.5 surpasses the Flash result with 58.6%. If this constitutes the key release for Google's return to prominence then shortcomings remain evident regarding coding ability. Anticipation builds for Gemini 3.5 Pro since that version will determine real capabilities.

3mo