gemini-4IMAGE

Gemini 4 Argon 五项 benchmark 分数流出

列出 Vals Index 68.9%、DeepSWE v1.1 77.9%、Vibe Code Bench 91.9%、RiemannBench 76.0%、GraphWalks 99.7%，称其在多数榜单领先 Opus 5.5、GPT-6 Astra 与 Fable 5.1。

Peng

@pengsonal28 天前收录原推发布于 2026.10.01

Peng @pengsonal原推发布于 2026.10.01

在 X 上打开原推

GOOGLE LAUNCHED GEMINI 4 ARGON AND IT'S ALREADY BEATING OPUS 5.5, GPT-6 ASTRA AND FABLE 5.1 the benchmark numbers are wild: • Vals Index → 68.9% • DeepSWE v1.1 → 77.9% • Vibe Code Bench → 91.9% • RiemannBench → 76.0% • GraphWalks → 99.7% it leads on most of the benchmarks shown here

这条是引用着另一条写的：x.com/google/status/2105388143902175529

x.com/pengsonal/status/2105394165731529012采集于 2026.10.01 10:01