Somebody posted this link earlier... Lots of history.
<<
http://www.emulators.com/docs/pentium_1.htm >>
Basic'ly, "x86" is a set of instructions. It's what and how the processor computes. Programs are written to use these instructions to do stuff.
The x86 instruction set began with the 8086. Then came the 80286, 80386, and 80486. Those are the actual chip numbers, and were shortened to 286, 386, and 486 for easy of use. The 386 is considered the base now, because it incorporated everything a modern OS system needs.
From there, Intel went to the Pentium, which would have been a 586, except you can't copyright a number. So, for marketing purposes, it was named Pentium. At the time, other chipmakers were producing 386s and 486s. Other chipmakers played the marketing game too, and produced a 6x86... It was a little less than the Pentium, but they tried to use the name to make people think it was better. It was not a 686.
Intel then made the first changes to the x86 instruction set and
started producing the Pentium Pro. This was the first 686 chip. Because Windows 95 was late, and thereby, Microsoft through in a bunch of 16-bit code. The PPro could only run 16-bit in emulation, and was a dog at it, and everybody blamed Intel. So, they produced the Pentlium MMX. It was just the 586 Pentium with some added instructions.
By that time, AMD was producing the K6 line... It was a 586, and a bit more than the Pentium, but less than the PPro or PMMX.
Intel went back to the drawing board and produced the Pentium II. It was based on the PPro, and thereby a 686. To try to out-market AMD, they played around with the first and second Celerons, which where just P2s with less cache. The onboard cace was the most expensive part of the chip, at the time. They shot themselves in the foot for awhile with the Celeron A when they put half the cache of the P2 on it, but ran it at full clock speed, rather than half clock speed like the P2. You could overclock the Celeron A and get almost the performance of a P2 for a fraction of the cost. That was the golden-age of overclocking.
AMD had gone back to the drawing board for 5 years to make the K7. This is the Athlons and Durons, and finally a chip to beat Intel. They are considered 686s, because they can run all the same instructions as a 686. Intel had added to the basic 386 instructions a couple times... MMX, MMX2, SSE, SSE2. Each minor set of intsructions was an optimization to the original set for some computer function. Mostly multimedia enhancements.
The P!!!, Celeron2, and P4 are all continuations of the P2 line. The Athlon MP and XP are continuations of the K7 line.
Next come the Itaniums and Athlon 64s. They are new chips. The A64s can run x86 instructions natively. This is good and bad news for the A64.. People can migrate to 64-bit instructions easier with the A64, but it also brings along all the bad parts of the x86 line. All code needs to be rewritten to 64-bit for the Itanium, but it leaves behind all the bad parts of the x86 line.