NewHow we catch a seller listing the wrong model, and refund you if one does.
UberLLM
deepseek logo

How do I use DeepSeek V3.2 in Claude Code?

deepseek/deepseek-v3.2

Claude Code reaches DeepSeek V3.2 through the Anthropic-compatible address of UberLLM. Set ANTHROPIC_BASE_URL to https://api.uberllm.dev (with no /v1 at the end), ANTHROPIC_AUTH_TOKEN to your UberLLM key and ANTHROPIC_MODEL to deepseek/deepseek-v3.2, then start claude. Set the opus, sonnet, haiku and subagent model variables to deepseek/deepseek-v3.2 as well: a slot left unset keeps asking for a Claude model at Claude prices. CLAUDE_CODE_MAX_CONTEXT_TOKENS is set to 160,000, the context size of DeepSeek V3.2, because Claude Code assumes 200,000 tokens for a model it does not know.

bash
export ANTHROPIC_BASE_URL=https://api.uberllm.dev
export ANTHROPIC_AUTH_TOKEN=ull_your_key
export ANTHROPIC_MODEL='deepseek/deepseek-v3.2'
export ANTHROPIC_DEFAULT_OPUS_MODEL='deepseek/deepseek-v3.2'
export ANTHROPIC_DEFAULT_SONNET_MODEL='deepseek/deepseek-v3.2'
export ANTHROPIC_DEFAULT_HAIKU_MODEL='deepseek/deepseek-v3.2'
export CLAUDE_CODE_SUBAGENT_MODEL='deepseek/deepseek-v3.2'
export CLAUDE_CODE_MAX_CONTEXT_TOKENS=160000
claude

Claude Code calls: Anthropic Messages API https://api.uberllm.dev/v1/messages · Full Claude Code guide

Replace ull_your_key with your UberLLM API key. One key works in every tool and for every model.

By API, DeepSeek V3.2 answers on three addresses with the same UberLLM key and the model id deepseek/deepseek-v3.2: OpenAI chat completions at https://api.uberllm.dev/v1/chat/completions, the OpenAI Responses API at https://api.uberllm.dev/v1/responses and the Anthropic Messages API at https://api.uberllm.dev/v1/messages.

How much does the DeepSeek V3.2 API cost?

On November 19, 2026, DeepSeek V3.2 costs $0.0026 per million input tokens and $0.0038 per million output tokens through UberLLM, 99% below its list price of $0.26 and $0.38. 92 sellers on the market can serve it. In the 7 days before, half of its 3,326 streamed requests got their first token within 5.4 seconds, with a typical prompt of 48,092 tokens. It reads up to 160,000 tokens of context. UberLLM adds no fee to these prices, and a request that fails is not billed.

Prices as of 09:00 UTC.

DeepSeek V3.2 in other tools

Other models in Claude Code