Models and prices
Open models with the refusals trained out. Every price is per million tokens and is exactly what your balance pays.
| Model | Context | Input / 1M | Output / 1M |
|---|---|---|---|
| Kimi K3.1 Prokimi-k3.1-pro-xploded | 1M | $7.50 | $37.50 |
| GLM 5.3glm-5.3-xploded | 1M | $2.00 | $5.00 |
| DeepSeek V4.1 Flash 1017deepseek-v4.1-flash-1017-xploded | 1M | $1.25 | $2.50 |
| Hy4hy4-xploded | 1M | $3.20 | $6.40 |
| Qwen3.8 27Bqwen3.8-27b-xploded | 1M | $1.50 | $5.00 |
| Qwen3.8 27B Ultrafastqwen3.8-27b-fast-xploded | 1M | $1.75 | $5.50 |
| MiMo V2.6 Flashmimo-v2.6-flash-xploded | 1M | $1.00 | $2.00 |
| GLM 5.3 Flashglm-5.3-flash-xploded | 1M | $0.30 | $1.00 |
| DeepSeek Flash Litedeepseek-v4.1-flash-lite-1017-xploded | 512K | $0.36 | $1.44 |
Cached input costs less on every model. A message is charged for the tokens it actually used, shown in your dashboard next to the reply.
At a glance
| Model | Refusals removed by | Images | Tools | Added |
|---|---|---|---|---|
| Kimi K3.1 Pro | Abliterated | Reads images | Calls tools | 6 Nov 2026 |
| GLM 5.3 | Uncensored build | Text only | Calls tools | Earlier |
| DeepSeek V4.1 Flash 1017 | Abliterated | Reads images | Calls tools | 24 Oct 2026 |
| Hy4 | Uncensored build | Reads images | Calls tools | 6 Nov 2026 |
| Qwen3.8 27B | Uncensored build | Reads images | Calls tools | Earlier |
| Qwen3.8 27B Ultrafast | Uncensored build | Reads images | Calls tools | 19 Oct 2026 |
| MiMo V2.6 Flash | Abliterated | Reads images | Calls tools | Earlier |
| GLM 5.3 Flash | Uncensored build | Reads images | Calls tools | Earlier |
| DeepSeek Flash Lite | Abliterated with Heretic | Reads images | Calls tools | 1 Nov 2026 |
Prices checked 12 November 2026.
Which one to pick
- Kimi K3.1 Pro is the strongest writer here and the dearest. Use it when quality matters more than price.
- GLM 5.3 is the all-rounder: long context, tool calling, steady at roleplay and code.
- DeepSeek V4.1 Flash 1017 is quick and cheap for long chats.
- GLM 5.3 Flash and DeepSeek Flash Lite are the cheapest, for high volume or testing.
More on each
- Kimi K3.1 Pro, uncensored→
- GLM 5.3, uncensored→
- DeepSeek V4.1 Flash, uncensored→
- Qwen3.8 27B, uncensored→
- The uncensored AI API→
- Set up SillyTavern→
- Set up opencode→
- Set up Claude Code→
- Abliterated vs uncensored→
- Cost calculator→
- Uncensored AI roleplay→
Using a model over the API
Send the id under the name as the model field. The API guide has full examples.