Core RankingsGrok 4.5
Core defaultby Grok
Core's default for strong general reasoning and agent work.
Intelligence54Artificial Analysis Index
Average task cost$0.50Observed benchmark task
Cost$2 in$6 out
Speed67tokens per second
Information
- Availability
- Available through Core's model router
- Core fit
- 59 / 100 · Strong
- Public intelligence
- 54 · high effort
- Benchmark snapshot
- Jul 2026
- Input price
- $2 per 1M tokens
- Output price
- $6 per 1M tokens
- Cache read
- $0.5 per 1M tokens