DeepSeek V4.1 Flash tops Tidelines usage ranking with 5.5T tokens/day, 22.6% share
maferase · reddit · 2026-10-14
Tidelines.ai's usage ranking shows DeepSeek V4.1 Flash leading with 22.64% share (5.5T tokens/day, $5.15M/day est.), larger than any single OpenAI, Anthropic or Google model. GLM 5.3 Flash (7.48%), Xiaomi MiMo-V2.6-Flash (7.16%), Tencent Hy4 preview (5.29%) and GPT-6 Luna (5.21%) round out the top five. Claude Opus 5.5 has only 3.35% share but the highest revenue at $12.4M/day. Chinese labs dominate the board with cheap, 1M-context models.
More from Infra
- The new case for a beefy PC: serving 5+ concurrent local AI users — AaronBergman18 · 2026-10-14
- Modal Trains RL on Mixed Accelerators, Cutting Cost While Matching Homogeneous Stability — charles_irl · 2026-10-14
- Ex-VC Katie Joins Modal's Fast-Growing NYC Office, Citing Talent Density and Brand Culture — charles_irl · 2026-10-14
- SurrealDB open-sources SurrealKV 1.0, a Rust storage engine 8x faster than RocksDB — mattturck · 2026-10-14
- Windows 11 flips CUDA device order and slows ComfyUI as Nvidia drops Win10 — BahBah1970 · 2026-10-14
- Musk: Micron's memory pricing is reasonable while rivals are 'extortionate' — elonmusk · 2026-10-14