gemini-4IMAGE

Leaked Gemini 4 Argon benchmarks: 68.9% knowledge work, 77.9% coding

Scores said to have leaked from a Google event show Argon leading Claude Opus 5.5 and GPT-6 Astra on both knowledge work and coding.

Haleemah

@HaleeeemahhArchived 20 days agoOriginally posted 2026.10.11

Leaked Gemini 4 Argon benchmarks: 68.9% knowledge work, 77.9% coding (image 1 of 2)

Leaked Gemini 4 Argon benchmarks: 68.9% knowledge work, 77.9% coding (image 2 of 2)

Haleemah @HaleeeemahhOriginally posted 2026.10.11

Open original on X

Gemini 4 argon benchmark scores from Google’s last event knowledge work: > gemini 4 argon - 68.9% > claude sonnet 5.5 - 67.0% > claude fable 5.1 - 65.8% > claude opus 5.5 - 67.0% coding: > gemini 4 argon - 77.9% > gpt-6 astra - 74.1% > claude fable 5.1 - 67.4% > claude opus 5.5 - 74.2%

x.com/haleeeemahh/status/2108954758388064282Archived 2026.10.12 09:51