deepseek-v4.1-flash-1017-xplodedDeepSeek V4.1 Flash 1017 is an uncensored Xploded build. Billing follows token usage; the same key works on both OpenAI- and Anthropic-compatible endpoints, and capabilities follow the model's live responses.
Pricing & specs
| Context | 1.05M (per the model provider's actual response) |
|---|---|
| Input price | $1.25 / million tokens |
| Cached input | $0.30 / million tokens (charged only when upstream usage reports a cache hit) |
| Output price | $2.50 / million tokens |
| Endpoints | OpenAI-compatible https://189.24.66.33/v1; Anthropic-compatible (Claude Code) https://189.24.66.33 |
| Billing | Settled on upstream-confirmed usage; requests without complete usage go to manual review |
| Modalities | Text + vision input |
| Reasoning | ✓ |
| Status | Online |
Call it with curl
curl https://189.24.66.33/v1/chat/completions \
-H 'authorization: Bearer sk-relay-...' \
-H 'content-type: application/json' \
-d '{"model":"deepseek-v4.1-flash-1017-xploded","messages":[{"role":"user","content":"hello"}]}'Call it with the OpenAI SDK
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://189.24.66.33/v1",
api_key="sk-relay-...",
)
resp = client.chat.completions.create(
model="deepseek-v4.1-flash-1017-xploded",
messages=[{"role": "user", "content": "hello"}],
)
print(resp.choices[0].message.content)Use it in Claude Code
Add the lines below to your shell config and run claude; use the model ID verbatim.
# Add to shell config, then run: claude export ANTHROPIC_BASE_URL=https://189.24.66.33 export ANTHROPIC_AUTH_TOKEN=sk-relay-... export ANTHROPIC_MODEL=deepseek-v4.1-flash-1017-xploded
Compare with the rest of the lineup
| Model | Context | Input / Mtok | Output / Mtok |
|---|---|---|---|
| GLM 5.3 Xploded | 1.05M | $2.00 | $5.00 |
| GLM 5.3 Flash Xploded | 1.05M | $0.30 | $1.00 |
| Kimi K3.1 Pro Xploded | 1.05M | $7.50 | $37.50 |
| Tencent Hy4 Xploded | 1.05M | $3.20 | $6.40 |
| DeepSeek V4.1 Flash 1017 Xploded | 1.05M | $1.25 | $2.50 |
| DeepSeek V4.1 Flash Lite 1017 Xploded | 512K | $0.36 | $1.44 |
| MiMo V2.6 Flash Xploded | 1.05M | $1.00 | $2.00 |
| Qwen3.8 27B Xploded | 1.05M | $1.50 | $5.00 |
| Qwen3.8 27B Ultrafast Xploded | 524K | $1.75 | $5.50 |
These are the final settlement rates; tool-calling and multimodal capabilities follow the model's actual responses.
FAQ
Is this model uncensored?
Xploded builds remove the base model's refusal training, so the refusal rate is far lower than the original; it is not a guarantee that every input gets answered — actual behavior follows the model's live responses.
Do you store my prompts?
No. Usage logs only record the model, token counts, cost, latency, status and trace ID — never prompts or response text.
How much does each call cost?
Input $1.25, output $2.50, cached input $0.30, all per million tokens; amounts are recorded at high precision.
Does it work with Claude Code / Cursor / Cline?
Yes. The same key works on both OpenAI- and Anthropic-compatible endpoints — just point your Base URL here. See the docs for details.