Most archaic IE language or macro-IE
Well, assuming Indo-European is valid, one of its most archaic language is clearly Hittite.
It is argued the wheel is an indo-european invention, and the word for "wheel" is known in Hittite, it is HURKIS.
Now, the word for "wheel" in Euskara (Basque) is GURPIL.
It is very easy to reconstruct KURKWITL, predecessor or HWEEL or WHEEL.
So we're left with those possibilities :
1 - Hittite wasn't Indo-European.
2 - Indo-European didn't invent the wheel, as Basque word GURPIL correlates with GURDI (cart).
3 - Proto-Basque and Hittite were related genetically.
Now, just take a look at Sanskrit KUMARA ("boy"), which is said to be pre-Indo-European substrate, well this is pure Basque KUME-ARA (child-male), "male child", the definition for "boy".
One interesting example is Latin CAVEMEN ("cavern, grotto, cave"), easily explained by GABE-MEN(DI) (hole-mountain), "hole in the mountain", the perfect definition for a cave. This is of interest since it traces a link back to paleolithic times, where our ancestors use to live in caves.
Now take a look at the numbers :
ZAZPI (Basque) and SISBI (Iberian) correlate very well with seven, sept (French) or sapta (Sanskrit), since SB does not exist in Indo-European, it makes sense it is linearized in ZAPIZ > ZAPTI (which can be nasalized in seven, no nasal sound in Basque united, but in some dialects they do exist).
You got the same with BéLTZ, as LTZ does not exist in Indo-European, the L glides after the B and you have old Norse BLATH and english BLACK. Just compare BELTZ KATUA and A BLACK CAT, you just have to put the A in the beginning in the sentence and this is copycat.
As for ZORTZI giving EIGHT or HUIT, ATTA etc... You simply have to delete the initial Z and you'll see it works perfectly.
BEDERATZI correlates to BERRI, as NINE is linked to NEW in every Indo-European language.
The link between FOUR and LAUR is easily explained with an initial KWAUR (with aspiration in middle KWA'UR, explaining Latin QUATUOR).
The same for HIRUR and THREE, by the way Zazaki (Persian) has the same form as Basque on this number. H was aspirated in Proto-Basque, so it is equivalent to THR.
BI(GA), the number TWO, is found as a noun initial meaning PARE / COUPLE in every Indo-European language.
Finally, the only oustiders left are ONE, FIVE and TEN.
But BAT have a different meaning in Basque, for the Indo-European cognate, juste take LEHENA(Z) (the first) : LEHENAZ > EENAZ > EINAS in Lituanian.
Now for BOSTEKOA giving FIVE, it requires more brain gymnastic, but is clearly feasible : BOSTEKOA > BEKWA > BEWA > FEWA.
HAMAR and TEN appear more problematic to reconcile, but in fact this works too : HAMAR > HAMR > HAN (pronounce THAN).
So we're left with total correspondence, and that can only make sense : Basque as a macro-Indo-European language, with a separation in Paleolithic times.
Or if you prefer, Basque was the first to branch off, before Hittite, before Armenian.