Deepseek V4 Pro via the Tokenator API
What the model is for
DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.
Specification
| Full name | Deepseek V4 Pro |
| API ID | deepseek-v4-pro |
| Model vendor | DeepSeek |
| Type | Text model (chat / completions) |
| Context window | 1.05M tokens |
| Max output | 393K tokens |
| Image input | no |
| Tokenator usage multiplier | 0.7× |
| Input modalities | text |
| Aliases | cursor-deepseek-v4-pro |
| Supported API formats | /v1/chat/completions /v1/responses /v1/messages |
| Upstream providers | 1 |
| Current status | available |
| Data updated | 2026-11-11 |
Benchmarks
| Intelligence index | 12.6 |
| Vals Index (Vals AI) | 38.6% |
| Code Migration (Vals AI) | 26.2% |
| IOI v1 (Vals AI) | 35.8% |
| LiveCodeBench (Vals AI) | 87.5% |
| ProgramBench (Vals AI) | 0% |
| SkillsBench (Vals AI) | 51.3% |
| SWE-bench (Vals AI) | 77.4% |
| Terminal-Bench 4.0 (Vals AI) | 11.1% |
| Vibe Code Bench v1.1 (Vals AI) | 49.9% |
| GPQA Diamond (Vals AI) | 89.4% |
| MMLU Pro (Vals AI) | 87.2% |
| ProofBench v1.1 (Vals AI) | 16% |
| CorpFin v2 (Vals AI) | 61.4% |
| EMB (Vals AI) | 51.6% |
| Finance Agent (v2) (Vals AI) | 44.1% |
| Tax Agent Bench (Vals AI) | 25.2% |
| TaxEval v2 (Vals AI) | 72.1% |
| CaseLaw v2 (Vals AI) | 59.4% |
| Harvey's Legal Agent Benchmark v1 (Vals AI) | 3.8% |
| Legal Research Bench (Vals AI) | 23.1% |
| LegalBench (Vals AI) | 80.3% |
| MedCode (Vals AI) | 40.5% |
| MedScribe (Vals AI) | 75.1% |
| Public Benefits Bench v1.1 (Vals AI) | 62.9% |
Artificial Analysis and Vals AI data loaded into the catalog for this model.
Upstream providers
| Provider | Multiplier |
|---|---|
| AiPartners Ltd | 0.7× |
A request goes to the first available provider by priority; if it fails, Tokenator switches to the next one.
Example request
curl https://api.tokenator.top/v1/chat/completions \ -H "Authorization: Bearer sk-your-tokenator-key" \ -H "Content-Type: application/json" \ -d '{ "model": "deepseek-v4-pro", "messages": [{"role": "user", "content": "Hello"}] }'
Compatible tools
The model can be used in any tool that accepts a custom base URL: Claude Code, Codex CLI, OpenCode, Cursor, Cline, Kilo Code, Cherry Studio, DeepSeek Harness, Hermes, Pi, OpenClaw.
See also
Comparisons involving this model: Claude Opus 4.6 vs Deepseek V4 Pro, Deepseek V4 Pro vs GLM 5.3 FlashX, Deepseek V4 Pro vs GPT 5.5, Deepseek V4 Pro vs Qwen 3.8 Max.
FAQ about Deepseek V4 Pro
What is the API ID of Deepseek V4 Pro?
deepseek-v4-pro — put this into the model field of your request.
What context window does Deepseek V4 Pro have?
1.05M tokens; max output is 393K tokens.
How many tokens will a request cost?
(input × 0.7×) + (output × 0.7×). See the documentation for details.
Which tools support this model?
Any tool that accepts a custom base URL: Claude Code, Codex CLI, OpenCode, Cursor, Cline, Kilo Code. Setup guides are in the integrations section.