Grok 4.7
xAI's frontier model for long-running agents, coding, knowledge work, and visual projects
Shogo Agent Index
Estimated#27 of 9254.8range 45.1–64.5 · 23% of weight measured
Agentic coding
58
26% measured
Long-horizon autonomy
69
30% measured
Knowledge-work agents
47
32% measured
Computer use
39
estimated
Security & context learning
50
estimated
Price
- Input
- $2.00/M
- Output
- $6.00/M
- Cache read
- $0.50/M
- 3:1 blend
- $3.00/M
- Agent-effective
- $0.98/M
Cheapest host for agents: Amazon Bedrock at $0.98/M.
Agent-effective = 20 input tokens per output token, 85% cache hits. Why.
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 80 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 77 | |
| DeepSWE | est. | 6% | predicted 53 | |
| FrontierSWE | 29.5% | 5% | proximus · $318.77/task | |
| CursorBench | 46.3% | 4% | source · $6.01/task | |
| GSO-Bench | est. | 4% | predicted 34 | |
| METR Time Horizons | est. | 10% | predicted 77 | |
| Vending-Bench 2 | $10,537 | 6% | source | |
| PostTrainBench | est. | 4% | predicted 26 | |
| APEX-Agents | 70.0% | 8% | source | |
| GDPval | est. | 7% | predicted 50 | |
| Remote Labor Index | est. | 5% | predicted 5 | |
| DeepResearch Bench | est. | 5% | predicted 49 | |
| OSWorld 2.0 | est. | 6% | predicted 12 | |
| OSWorld | est. | 4% | predicted 80 | |
| Cybench | est. | 5% | predicted 78 | |
| CL-bench | est. | 5% | predicted 23 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| ARC-AGI | Reasoning | 90.2% | source |
| ARC-AGI-2 | Reasoning | 61.4% | source |
| Chess Puzzles | Reasoning | 41.0% | source |
| CritPt | Reasoning | 18.0% | source |
| DTBench | Knowledge | 96.0% | source |
| FrontierCode | Coding | 47.6% | grok-build |
| FrontierMath Tiers 1-3 | Math | 53.0% | source |
| FrontierMath Tier 4 | Math | 17.1% | source |
| GeoBench | Multimodal | 61.0% | source |
| GPQA Diamond | Reasoning | 92.7% | source |
| LMCA | Knowledge | 49.4% | source |
| MirrorCode | Coding | 54.4% | source |
| OTIS Mock AIME | Math | 98.1% | source |
| ProofBench | Math | 26.0% | source |
| SciCode | Coding | 57.8% | source |
| SimpleQA Verified | Knowledge | 56.0% | source |
| WebDev Arena | Coding | 1638 | source |
| WeirdML v3 | Coding | 8.5% | codex_cli |
API hosts
Live prices and uptime for every host serving Grok 4.7 via OpenRouter.
| Host | Input | Output | Cache read | Agent $/M | Uptime 24h | Quant |
|---|---|---|---|---|---|---|
| Amazon Bedrock | $2.00 | $6.00 | $0.50 | $0.98 | 98.7% | — |
| xAI | $2.00 | $6.00 | $0.50 | $0.98 | 93.4% | — |
| xAI | $2.00 | $6.00 | $0.50 | $0.98 | 97.6% | — |
| xAI | $4.00 | $12.00 | $1.00 | $1.95 | 96.3% | — |
| xAI | $4.00 | $12.00 | $1.00 | $1.95 | 100.0% | — |