So I took a picture at a high ISO value in poor light conditions. I can't retake it, the grain effect is bad, I'm aware this will never be a good photo, but I still want to make the best out of it. So, photo-editing software and... then what?
What filters/tools/techniques to apply to maximally reduce the ISO noise and restore the image to maximum clarity?
GIMP preferred, though if you know good techniques that work with other open source software, please share too.
I'm not sure about GIMP, I use Lightroom 4 and there is this wonderful plugin Noiseware 5. I used it myself and I got good results. Some report Noise Ninja to be a good tool too but I didn't try it myself.
I'm sure that you can do this in GIMP without plugins, a quick googling results that maybe helpful:
Adding some info to Akram's answer.
You could use try some of the filters from G'mic (a "plugin pack" for the Gimp which has a lot more filters than just noise removal ones). Here's its website and download page.
For a tutorial: Noise reduction with G'Mic . An excerpt:
Anisotropic Smoothing is the best solution for pure noise reduction, it can be found under Enhancement and will give you the lower loss of details, and last but not least is really easy to setup. [...] Generally you should have really good results with default values, if you need more sharpness, change the value from 0.70 to 0.90.
When I'm in a hurry I personally use Iain's Fast Denoise because it's faster.
Usually, the blue channel has more noise than the others. So an old trick is applying a slight blur to the blue channel might help.
That said, noise reduction software like Noiseware and Noise Ninja are so far advanced beyond this old trick that you might find it worth your time and money to look into a more automated solution.
The Wavelet denoise GIMP Plugin might be interesting for you. See for example this video tutorial and Post processing on low light/fast shutter photo.
I think RawTherapee also has wavelet tools (I'm no expert, though), see http://scribble-jpc.blogspot.se/2015/03/first-view-wavelet-tool.html and http://scribble-jpc.blogspot.co.uk/2012/07/first-look-rawtherapees-new-noise.html in case you are looking for any open-source solution.
You can always scale down your picture to reduce noise. The scaling operation will average groups of noisy pixels, this cancels some of the noise out. You can do this alone or after you apply other noise reduction techniques.
