I've been thinking about this a bit too primarily b/c of the thread about case studies in science talent development. The article linked there seems to define "gifted" as the intersection btwn high IQ and high creativity. Nowhere does that necessitate high achievement, though, since one can be of high IQ/ability and not performing highly in school for a variety of reasons.

Page 37 has the following table (formatting is messed up here I'm sure):

Table 1. Means and Standard Deviations for IQ and Creativity Test Scores (Z Scores for the Four Ability Groups: Gifted, Creative, Intelligent, and Average)

IQ test Creativity test
Group n M SD M SD
Average 12 99.83 4.63 49.67 5.02
Gifted 11 129.82 4.60 64.00 2.93
Creative 11 106.18 8.60 66.54 3.72
Intelligent 15 127.33 3.75 48.87 4.78

In other words, they state that:

The "average" kids had:
average IQs and average creativity

The "gifted" kids had:
IQs around the 98th percentile plus or minus a bit and high creativity

The "creative" kids had:
average IQs and high creativity

The "intelligent" kids had:
IQs pretty close to as high as the gifted kids and average creativity