GitHub[150K]Try OpenCode

DeepSeekdeepseek/deepseek-v4-pro

DeepSeek V4 Pro

Ranked #2 across last week's OpenCode Go usage with 27% of observed 2M volume.

Model weights: Hugging Face

7D Rank#2

Even vs previous week

Overview. Recent tokens, sessions, and market position.

Tokens20T

last two months

Sessions3,376,626

completed sessions

Token Share27%

19 models

Momentum+100%

vs previous window

Usage. Daily token volume over the recent two-month window.

0APR 21

0APR 22

0APR 23

10BAPR 24

49BAPR 25

70BAPR 26

118BAPR 27

149BAPR 28

161BAPR 29

163BAPR 30

167BMAY 1

196BMAY 2

219BMAY 3

287BMAY 4

295BMAY 5

304BMAY 6

306BMAY 7

287BMAY 8

276BMAY 9

253BMAY 10

361BMAY 11

384BMAY 12

369BMAY 13

339BMAY 14

361BMAY 15

278BMAY 16

282BMAY 17

362BMAY 18

383BMAY 19

327BMAY 20

368BMAY 21

434BMAY 22

428BMAY 23

409BMAY 24

588BMAY 25

612BMAY 26

569BMAY 27

500BMAY 28

492BMAY 29

393BMAY 30

395BMAY 31

534BJUN 1

573BJUN 2

565BJUN 3

511BJUN 4

469BJUN 5

403BJUN 6

415BJUN 7

615BJUN 8

658BJUN 9

630BJUN 10

622BJUN 11

570BJUN 12

446BJUN 13

428BJUN 14

441BJUN 15

Efficiency. Cost, cache behavior, and average session shape.

Cost$1.5M

total spend

Cost / 1M$0.43 / $0.87

input / output

Cost / Session$0.45

average

Tokens / Session5.9M

average

Cache Ratio97%

input tokens

Geo Breakdown. Model tokens used by country.

#01China

1.6T17%

Peers. Nearby models by recent token volume.