BENCHMARK
A score only means something
inside a concrete scenario
The three metrics in the table below come from one batch of tasks: the same sketch, the same voice-over script, the same repository.
01 / MODEL & COST
Model capability × price board
Capability scores and price per million tokens for 0 models on the same batch of tasks. The x axis is logarithmic because the cheapest and the most expensive are two orders of magnitude apart. Prices are in CNY per million tokens.
Model
Capability
¥/M tok
The further to the upper left, the better the value. Context length and throughput are in the table and the hover tips.
02 / HEAD TO HEAD
Leaders of three axes, compared
03 / OVERALL
Overall ranking · top ten
No fully scored entries yet.