Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick — Balanced (price + speed), Nitro (fastest), or Exacto (highest tool-calling accuracy).
The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.
Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).
Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.
Scores on standardized evaluations. Higher percentages are better — and rank percentile shows where this model lands among all models on OpenRouter.
| Source | Benchmark | Score |
|---|---|---|
| Artificial Analysis | Grok 4.5 (high) Intelligence Index | 39.1 |
| Artificial Analysis | Grok 4.5 (high) Coding Index | 72.4 |
| Artificial Analysis | Grok 4.5 (high) Agentic Index | 42.1 |
| Artificial Analysis | Grok 4.5 (high) GPQA Diamond | 93.1% |
| Artificial Analysis | Grok 4.5 (high) HLE | 42.7% |
| Artificial Analysis | Grok 4.5 (high) AA-LCR | 79.3% |
| Artificial Analysis | Grok 4.5 (high) GDPval-AA | 46.5% |
| Artificial Analysis | Grok 4.5 (high) CritPt | 15.4% |
| Artificial Analysis | Grok 4.5 (high) SciCode | 55.0% |
| Artificial Analysis | Grok 4.5 (high) AA-Omniscience Accuracy | 51.5% |
| Artificial Analysis | Grok 4.5 (high) AA-Omniscience Non-Hallucination Rate | 45.9% |
| Design Arena | Grok 4.5 Agents Arena Agenticgamedev Elo | 1235 |
| Design Arena | Grok 4.5 Agents Arena Androidnative Elo | 1247 |
| Design Arena | Grok 4.5 Agents Arena Full Stack Elo | 1261 |
| Design Arena | Grok 4.5 Agents Arena Godotgamedev Elo | 1270 |
| Design Arena | Grok 4.5 Agents Arena Htmlslides Elo | 1208 |
| Design Arena | Grok 4.5 Agents Arena Mobile Apps Elo | 1244 |
| Design Arena | Grok 4.5 Agents Arena Python-Pptxslides Elo | 1230 |
| Design Arena | Grok 4.5 Agents Arena Webapps Elo | 1231 |
| Design Arena | Grok 4.5 Models Arena 3D Elo | 1288 |
| Design Arena | Grok 4.5 Models Arena Asciiart Elo | 1270 |
| Design Arena | Grok 4.5 Models Arena Code Categories Elo | 1292 |
| Design Arena | Grok 4.5 Models Arena Data Visualization Elo | 1286 |
| Design Arena | Grok 4.5 Models Arena Game Development Elo | 1288 |
| Design Arena | Grok 4.5 Models Arena SVG Elo | 1238 |
| Design Arena | Grok 4.5 Models Arena UI Component Elo | 1293 |
| Design Arena | Grok 4.5 Models Arena Website Elo | 1296 |
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.
Token volume and request traffic to this model over time.
Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.
Grok 4.5 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM.
Grok 4.5 costs $2.00/M input tokens and $6.00/M output tokens, with separate rates for Cache Read at $0.30/M tokens and Web Search at $5.00/1K calls.
Grok 4.5 has a 500,000 token context window.
Yes. Grok 4.5 accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
Grok 4.5 accepts text, images, and files such as PDFs as input and returns text.
Grok 4.6, Grok Build 0.1, Grok 4.3 and 2 more are other text models from SpaceXAI.
Grok 4.5 was released on July 8, 2026.