deepseek/deepseek-v3.2-exp
DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| Novitafp8 | $0.270 | $0.410 | 164K | 99.9% |
| SiliconFlowfp8 | $0.270 | $0.410 | 164K | 99.4% |
| AtlasCloudfp8 | $0.270 | $0.410 | 164K | 100% |
DeepSeek V3.2 Exp costs $0.270 per million input tokens and $0.410 per million output tokens via OpenRouter, making it 123rd cheapest of 298 paid models.
DeepSeek V3.2 Exp scores 32.1 on the Artificial Analysis Intelligence Index, ranking 69th of 180 benchmarked models, with a GPQA Diamond score of 75%.
DeepSeek V3.2 Exp generates around 18 tokens per second with 1.2s time-to-first-token (p50), the 275th fastest tracked model.
DeepSeek V3.2 Exp supports a 164K-token context window and can output up to 66K tokens. It accepts text input.