Grok 4.5 (high) API Pricing
SpaceXAI · released 2026-07-08
Pricing
0.200¢ / 1K
0.600¢ / 1K
0.300¢ / 1K
What it actually costs
| Workload | Tokens | Cost |
|---|---|---|
| Short chat reply | 500 in / 300 out | 0.28¢ |
| RAG query | 4,000 in / 500 out | $0.0110 |
| Long document summary | 50,000 in / 1,000 out | $0.1060 |
| Code review of a 2k-line file | 30,000 in / 2,000 out | $0.0720 |
| 1,000 chat replies/day for 30 days | 30,000 × (500 in / 300 out) | $84.00 |
Cost = (input tokens ÷ 1,000,000 × $2.00) + (output tokens ÷ 1,000,000 × $6.00).
Benchmarks
#26 of 533
#17 of 197
Similar intelligence, lower price
| Model | Intelligence | Blended $/1M |
|---|---|---|
| Qwen3.8-Flash-Next | 39.8 | $0.23 |
| GLM 5.3 Flash | 41.8 | $0.237 |
| GPT-5.6 Luna (max) | 37.3 | $0.45 |
| DeepSeek V4.1 Flash (Reasoning, Max Effort) | 39.5 | $0.525 |
| Gemini 3.8 Flash (high) | 40.9 | $1.50 |
Leaderboard neighbours
| Model | Intelligence | Blended $/1M |
|---|---|---|
| Muse Spark 1.2 (xhigh) | 39.6 | $2.00 |
| DeepSeek V4.1 Flash (Reasoning, Max Effort) | 39.5 | $0.525 |
| GPT-5.4 (xhigh) | 39.0 | $5.625 |
| Grok 4.5 (high) | 38.8 | $3.00 |
| GPT-5.5 (xhigh) | 38.4 | $11.25 |
| Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 38.2 | $4.00 |
| GPT-5.6 Luna (max) | 37.3 | $0.45 |
More from SpaceXAI
| Model | Intelligence | Blended $/1M |
|---|---|---|
| Grok 4.7 (xhigh) | 46.4 | $3.00 |
| Grok 4.6 (high) | 44.3 | $3.00 |
| Grok Build 0.1 0616 | 27.2 | $1.25 |
| Grok 4.20 0309 v2 (Reasoning) | 25.7 | $1.563 |
| Grok 4.20 0309 (Reasoning) | 25.2 | $3.00 |
| Grok 4.3 (high) | 24.9 | $1.563 |
| Grok 4 | 22.5 | $6.00 |
| Grok 4.1 Fast (Reasoning) | 20.4 | — |
Count tokens & estimate cost for Grok 4.5 (high) See Grok 4.5 (high) on the leaderboard
Source: Artificial Analysis · last synced 2026-09-22
FAQ
Who develops Grok 4.5 (high)?
Grok 4.5 (high) is developed by SpaceXAI, released 2026-07-08.
How much does Grok 4.5 (high) cost per million tokens?
Grok 4.5 (high) costs $2.00 per 1M input tokens and $6.00 per 1M output tokens ($3.00 blended), per Artificial Analysis's official pricing data.
How does Grok 4.5 (high) rank on benchmarks?
Grok 4.5 (high) scores 38.8 on the Artificial Analysis Intelligence Index, ranking #26 of 533 models we track.
Is there a cheaper model with similar intelligence to Grok 4.5 (high)?
Yes — Qwen3.8-Flash-Next scores similar intelligence (39.8) at $0.23/1M blended, versus $3.00/1M for Grok 4.5 (high).
Where does this data come from?
Pricing and benchmark data for Grok 4.5 (high) comes from Artificial Analysis (artificialanalysis.ai), synced daily. Last synced 2026-09-22.