Repository navigation
Every LithosAI model, not just one
The LithosAI card used to list only DeepSeek V4.1 Flash's three tiers. LithosAI sells four models, so each one now gets its own line:
| Model | From | Speed |
|---|---|---|
| Kimi K3 | ×10 |
80+ tok/s |
| DeepSeek V4.1 Flash | ×0.50 |
250+ tok/s |
| GLM-5.3 | ×2.75 |
200+ tok/s |
| GLM-5.3-Flash | ×0.42 |
250+ tok/s |
A line shows the cheapest tier on sale, its multiple of DeepSeek's peak output price (×1), and the single-user speed LithosAI advertises. Hover a line and a bubble drops down with the rest of that model's tiers — every tier's price and speed, in the same columns as the card above, with the model name still visible. GLM-5.3 only sells Base, GLM-5.3-Flash only Base and Ultra, so neither has more to show.
Qwen 3.8, Gemma 4 and GPT OSS are not priced on LithosAI's page yet, so they are not listed.
Also in 1.6
- The bubble never covers its own row and always drops in the same direction; rows it covers step out of the way instead of being sliced mid-line.
- README and screenshots refreshed for 1.6.
Full Changelog: v1.5...v1.6