F₁₇ = 1597 and F₁₈ = 2584, and the workbench puts their quotient at 1.6180338134001. φ itself is 1.6180339887499, so the two agree through 1.618033 and then part company — the convergent is the smaller of the two, and the third link prints the gap as −1.7534976959332⋅10⁻⁷. Consecutive Fibonacci ratios close in on φ from alternate sides, each one better than the last, and eighteenth is where the sixth decimal finally settles.
The reason these particular fractions are the best there is comes from the continued fraction. φ = [1; 1, 1, 1, …], all ones, which is the slowest-converging continued fraction in existence: no irrational number is harder to approximate by fractions, since a larger partial quotient somewhere would hand you an unusually good rational for free. Its convergents are exactly the ratios Fₙ₊₁/Fₙ, which is the same fact Kepler noticed, stated in the language of place value instead of sequences.
Answer
1.6180338134001
How to read it
2584/1597 on the workbench, fourteen displayed digits of an irrational-looking decimal that is actually rational. The exact φ begins 1.61803398875.
Source
Kepler noted in 1611 that the ratio of consecutive Fibonacci numbers approaches the golden ratio; the sequence itself is from Liber Abaci (1202)
Work it in the calculator
These links open the right calculator with this problem's numbers already typed in. The line above is what the screen should show — check it yourself.