Models
The following models are used in the Free / Standard plan rotation pool. Weights reflect relative traffic allocation.
| Model | Provider | Weight | Input ($/1M) | Output ($/1M) | Total ($/1M)† |
|---|---|---|---|---|---|
bedrock:openai-gpt-120b-oss | AWS Bedrock (OpenAI) | 40 | $0.15 | $0.60 | $0.75 |
bedrock:kimi-k2.5 | AWS Bedrock (Moonshot AI) | 28 | $0.60 | $3.00 | $3.60 |
claude-haiku-4.5 | Anthropic | 18 | $1.00 | $5.00 | $6.00 |
bedrock:glm-5 | AWS Bedrock (Zhipu AI) | 15 | $1.00 | $3.20 | $4.20 |
gpt-4o-mini | OpenAI | 10 | $0.15 | $0.60 | $0.75 |
gpt-5.4-mini | OpenAI | 7 | $0.75 | $4.50 | $5.25 |
deepseek-v4-flash | DeepSeek | 5 | $0.14 | $0.28 | $0.42 |
bedrock:qwen3-vl-235b | AWS Bedrock (Alibaba) | 5 | $0.22 | $0.88 | $1.10 |
† Total = Input + Output rates combined (per 1M tokens each).
Bring Your Own Key (BYOK): Users who supply their own API key can select any of the models listed above — including premium models not included in the default rotation — and are billed at the provider's standard rates.