This is my 2006 with 272k miles. I don't talk about it much. Got it in January with a bad catalytic converter. Turns out selling an OE cat more than pays for an aftermarket replacement.
Anywho, nothing was up with the battery until recently. I'd depart for work in the morning and the car would go from "80%" to "one bar" in 1/2 mile. The car by design idles to warm up unless you really romp on it, making up for the difference with electrical power. Mine would stumble as I pressed the accellerator, as if there wasn't the electricity available the car would have expected. I ran "torque" with the 2nd gen extra PIDs and noticed mostly even battery voltages. Finally, it threw code P0A80 "replace HV battery" but without any additional codes hinting at bad modules. I cleaned the battery cooling fan and fixed its faulty relay connection, to no improvement in performance. It gets (got?) 48-49 MPG vs my 2005 which gets 53-54 under the same conditions.
I tried watching the voltage sags when I gunned it and noticed momentary blips in the #2 block pair-of-modules. (Toyota monitors 14 blocks for 28 modules or 168 NiMH cells all wired in series.)
So I diagnosed a module or modules that have no shorted cells but reduced capacity, and had hopes it was #2.
I rolled the car into the garage on "three bars", nearly empty, and disassembled the pack. Used 2x 55watt fog lights to load a pair of modules (~15 volts). Took readings before turning the lights on, and after 5 minutes with them still on. Right off the bat I found a more substantial sag in #2. Went after the modules individually and one was normal but the other down a volt. Marked it with sharpie to remember it's a dud.
Worked my way down the line, found no others bad. Reassembled with two pre-owned modules, and ran them under the same test as the rest. (I wanted them balanced, and had no great way of recharging the "good" single in #2.) Passed the test and they were all within 1/10 volt at rest.
Reassembled. Car assumed I had 70 percent charge as some default. This was reevaluated by the computer within 30 seconds and I had one "pink bar". Torque reported 35% SOC.
Drove the car four miles, and the battery's up to 65% SOC now, and 7 blue-bars. This is actually great news. Under the old module it would go from nothing to "100% full" in very little time. Under normal conditions it should only get 100% full after a long downhill descent with regenerative breaking.
On to the pictures. Don't kill yourself working on these things, kids, there's 230 Volts DC if you try to find it, but it's safe if you take precautions.