OpenAI: GPT-5 vs Grok 4 Fast
Grok 4 Fast is cheaper than OpenAI: GPT-5 by about 92% on a blended (75% input / 25% output) basis. OpenAI: GPT-5 charges $1.250000 / $10.000000 per 1M input/output tokens; Grok 4 Fast charges $0.200000 / $0.500000.
| Field | OpenAI: GPT-5 | Grok 4 Fast |
|---|---|---|
| Input $/Mtok | $1.250000 | $0.200000 |
| Output $/Mtok | $10.000000 | $0.500000 |
| Cache read $/Mtok | $0.125000 | – |
| Cache write $/Mtok | – | – |
| Input $/Mchar | 0.352142 | n/a |
| Output $/Mchar | 2.817139 | n/a |
| Context window | 400000 | 2000000 |
| Tokenizer | o200k | – |
Price history
Same workload, monthly cost
Illustrative monthly cost at 10M tokens/month, split per workload preset. Not a substitute for the calculator, which uses your own volume, cache and batch assumptions.
| Workload | OpenAI: GPT-5 | Grok 4 Fast |
|---|---|---|
| Coding agent | $16.88 | $2.15 |
| RAG chat | $25.63 | $2.45 |
| Summarisation (batch) | $21.25 | $2.30 |
| Classification | $14.25 | $2.06 |
| Chat assistant | $47.50 | $3.20 |
| Content generation | $82.50 | $4.40 |
Cite this page
OpenAI: GPT-5 vs Grok 4 Fast: $1.250000/$10.000000 vs $0.200000/$0.500000 per 1M tokens, as of 2026-10-04 (Token Barometer, https://tokenbarometer.com/compare/openai_gpt-5-vs-xai_grok-4-fast).