DeepSeek
Direct inference API from DeepSeek. Model aliases can point to specific current model releases.
Models on DeepSeek 2
A curated selection, with links to other providers for the same model.
Explore imported models by type. Image and video categories refer to generation; image understanding is available under Vision.
Artificial Analysis Intelligence Index ↗ · Higher is better · Checked 2026-10-02
Rank among verified models in this catalog. Best published configuration per model; equal scores share a rank. Scores are rounded by the source. Unverified and non-language models appear last without a rank. Methodology ↗2 models
| Rank | Model | Intelligence | Best price | Input → output | Providers |
|---|---|---|---|---|---|
| #20 | DeepSeek V4.1 FlashDeepseek | 39max | $0.15 in / $0.6 outUSD / 1M tokensDeepSeek Off-peak · Lowest listed input price; output is from the same tariff. Time-of-day tariff; consult official source for UTC peak/off-peak windows and alias behavior. Compare prices → | text, image → text | DeepInfraDeepSeekFireworks AIOpenRouterTogether AI |
| #23 | DeepSeek V4 Pro 0813Deepseek | 36max | $0.22 in / $4.2 outUSD / 1M tokensOpenRouter Router catalog · Lowest listed input price; output is from the same tariff. OpenRouter route catalog price, not a direct-provider price. Routing, modality and endpoint conditions can change the actual bill. Compare prices → | text → text | DeepSeekNovita AIOpenRouter |
Catalog entries link to their specification and availability sources. This is a curated selection, not every model each provider offers. Best prices are the lowest collected USD tariffs, not a total-cost estimate. Token offers are selected by input price, with output from the same tariff; workload mix may change the cheapest provider. Listed variants may require batch processing, off-peak hours, or specific media settings. Price coverage varies by provider; collected tariffs include their source and observation time. Benchmark scores are a dated snapshot, not a live feed or a universal measure of quality. They apply to the shown evaluation configuration; listed prices may use different settings. Models without verified scores are not assumed to be weaker.