Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work
DeepSeek V4 Flash Vision Exp is offered by 17 providers on this page. Its official list price is $0.15 per 1M input tokens and $0.60 per 1M output tokens. The lowest paid channel is CrofAI at $0.08 / $0.20 per 1M, about 5.5× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it.
The context window is 1,000,000 tokens at the reference host, but hosts report different limits, from 1,000,000 to 1,048,576, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text and Image. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Its training knowledge cuts off at 2025-05.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| CrofAI | Gateway | $0.08 | $0.20 | $0.007 | — | 1,000,000 | 131,072 | |
| OpenCode Zen | Gateway | $0.14 | $0.28 | $0.028 | — | 1,000,000 | 384,000 | |
| AMD DeepSeek-V4-Flash-Vision-Exp | Cloud | $0.14 | $0.28 | $0.0028 | — | 1,000,000 | 384,000 | |
| OrcaRouter | Gateway | $0.15 | $0.30 | $0.02 | — | 1,000,000 | 384,000 | |
| DeepSeekOfficial | First-party | $0.15 | $0.60 | $0.003 | — | 1,000,000 | 393,216 | deprecated |
| OpenCode Go | Gateway | $0.15 | $0.60 | $0.003 | — | 1,000,000 | 384,000 | |
| Eden AI | Gateway | $0.15 | $0.60 | $0.003 | — | 1,000,000 | 384,000 | |
| AIHubMix | Gateway | $0.16 | $0.62 | $0.0031 | — | 1,000,000 | 384,000 | |
| OpenRouter | Gateway | $0.22 | $0.65 | $0.0069 | — | 1,048,576 ⚠ | 262,144 | |
| Vercel AI Gateway | Cloud | $0.22 | $0.65 | $0.0068 | — | 1,048,576 ⚠ | 1,048,576 | |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 7 of 17 providers.
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| Terminal-Bench | harness: DeepSeek Harness minimal mode · variant: max effort · v2.1 | 83.9 | score | Source ↗ |
| NL2Repo | harness: DeepSeek Harness minimal mode · variant: max effort | 57.7 | score | Source ↗ |
| CyberGym | harness: DeepSeek Harness minimal mode · variant: max effort | 75.3 | score | Source ↗ |
| DeepSWE | harness: DeepSeek Harness minimal mode · variant: max effort | 59.3 | score | Source ↗ |
| Toolathlon-Verified | harness: DeepSeek Harness minimal mode · variant: max effort | 75.9 | score | Source ↗ |
| DSBench-Hard | harness: DeepSeek Harness minimal mode · variant: max effort · dataset: internal | 63.6 | score | Source ↗ |
| AutomationBench | harness: DeepSeek Harness minimal mode · variant: max effort · dataset: public | 25.7 | score | Source ↗ |
| ApexBench | — | 36.5 | pass@1 | Source ↗ |
| Agents' Last Exam | — | 27.3 | score | Source ↗ |
| Chartography | — | 64.3 | score | Source ↗ |
| ZeroBench | — | 35 | pass@5 | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.