Compare models
Pick two to four models. No markup on model usage: you pay the provider's list price.
| Attribute |  Gemini 3.8 Flash |   DeepSeek V4 Pro |
|---|---|---|
| Provider | DeepSeek | |
| Model id | google/gemini-3.8-flash | deepseek/deepseek-v4-pro |
| Status | Available via Safeguard | Available via Safeguard |
| Input / 1M tokens | $0.75 | $1.32 |
| Output / 1M tokens | $3.75 | $3.96 |
| Price notes | Paid tier. $0.75 in / $3.75 out (output includes thinking tokens) through 2026-12-31, rises to $1.50 / $7.50 from 2027-01-01. | Peak-hour list price (cache miss). Off-peak is half: $0.66 in / $1.98 out. Cache-hit input $0.044 peak. DeepSeek says it is phasing V4 Pro out but continues to serve it. |
| Context | 1.05M | 1M |
| Parameters | Not published | Not published |
| Max output | 65,536 tokens | 384,000 tokens |
| Modality | Text and image | Text |
| Weights | Closed | Open |
| Zero data retention | Not offered | Not stated |
| Prompts go to | Google (Gemini Developer API) | DeepSeek (People's Republic of China) |
| Retention | Paid tier: prompts and responses not used to improve Google products, but logged for a limited period for abuse monitoring, free tier content may be used to improve products | Privacy policy: personal data stored in the PRC and kept as long as necessary to provide the service, may be used to train models, with a right to opt out. No API-specific retention period or ZDR offering published |
| Region | No region guarantee, data may be stored transiently or cached in any country where Google operates. Google points ZDR and enterprise processing needs to Vertex AI (Gemini Enterprise Agent Platform) | China (PRC) |
| Price checked | 2026-10-10 | 2026-10-10 |
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.