Grok 4.5 via the Tokenator API
What the model is for
Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.
Specification
| Full name | Grok 4.5 |
| API ID | grok-4.5 |
| Model vendor | xAI |
| Type | Text model (chat / completions) |
| Context window | 500K tokens |
| Max output | 128K tokens |
| Image input | yes |
| Tokenator usage multiplier | 1.7× |
| Input modalities | text, image, file |
| Aliases | cursor-grok-4.5 |
| Supported API formats | /v1/chat/completions /v1/responses /v1/messages |
| Upstream providers | 1 |
| Current status | available |
| Data updated | 2026-11-11 |
Benchmarks
| Coding index | 72.4 |
| Agentic index | 41.2 |
| Vals Index (Vals AI) | 44.7% |
| Vals Multimodal Index (Vals AI) | 63.4% |
| Vals Web Search Index (Vals AI) | 35.5% |
| Code Migration (Vals AI) | 36.6% |
| IOI (Vals AI) | 40.6% |
| LiveCodeBench (Vals AI) | 87.4% |
| ProgramBench (Vals AI) | 0% |
| SkillsBench (Vals AI) | 66% |
| SWE-bench (Vals AI) | 86.6% |
| Terminal-Bench 4.0 (Vals AI) | 8.6% |
| Vibe Code Bench v1.1 (Vals AI) | 69% |
| GPQA Diamond (Vals AI) | 92.9% |
| MMLU Pro (Vals AI) | 89.2% |
| MMMU Pro (Vals AI) | 61.8% |
| ProofBench v1.1 (Vals AI) | 31% |
| CorpFin v2 (Vals AI) | 67.4% |
| EMB (Vals AI) | 52.9% |
| Finance Agent (v2) (Vals AI) | 48.3% |
| MortgageTax (Vals AI) | 61.8% |
| Tax Agent Bench (Vals AI) | 24.9% |
| TaxEval v2 (Vals AI) | 71.7% |
| Harvey's Legal Agent Benchmark v1 (Vals AI) | 12.9% |
| Legal Research Bench (Vals AI) | 38% |
| LegalBench (Vals AI) | 86% |
| MedCode (Vals AI) | 43.3% |
| MedScribe (Vals AI) | 86.9% |
| CyberBench v1.1 (Vals AI) | 71% |
| SRE Bench (Vals AI) | 0.8% |
| SAGE v1 (Vals AI) | 35% |
Artificial Analysis and Vals AI data loaded into the catalog for this model.
Upstream providers
| Provider | Multiplier |
|---|---|
| AiPartners Ltd | 1.7× |
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": "grok-4.5", "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: GPT 5.5 vs Grok 4.5.
FAQ about Grok 4.5
What is the API ID of Grok 4.5?
grok-4.5 — put this into the model field of your request.
What context window does Grok 4.5 have?
500K tokens; max output is 128K tokens.
How many tokens will a request cost?
(input × 1.7×) + (output × 1.7×). 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.