SpaceXAIGrok 4.7
Grok 4.7 is a SpaceXAI proprietary reasoning model, released 21 Sept 2026. Its best configuration (high reasoning effort) ranks #33 of 431 on the BenchLeader Index at 65.8 ±8.6, in the upper half. It scores highest in composite (88) and lowest in long context (65). At $3.00 per million tokens blended it is among the most expensive ranked models. Output speed of 55 tokens per second puts it slower than most, with a first token in 1.4 s. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 21 Sept 2026.
- Blended price
- $3.00/M
- $2.00 in · $6.00 out
- Output speed
- 55 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 1.41 s
- Context
- 500k
- Full answer
- –
- Cost per run
- –
- Released
- 21 Sept 2026
- Overall index66
- Reasoning76
- Knowledge73
- Long context65
- Composite88
Versions
SpaceXAI has shipped 9 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| Humanity's Last Exam (AA)not in index | 42.3%#43 | 43.1%#34 | Artificial Analysis |
| CritPt | 18.0%#44 | 17.7%#46 | Artificial Analysis |
| MysteryMechanismnot in index | – | 25.7%#8 | Vals AI |
Coding
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| SciCode (AA)not in index | 57.8%#13 | 57.4%#15 | Artificial Analysis |
| IOI | – | 39.4%#21 | Vals AI |
| Code Migrationnot in index | 34.7%#24 | – | Vals AI |
| ProgramBenchnot in index | 0.0%#16 | – | Vals AI |
| Vibe Code Bench v1.1not in index | – | 75.9%#19 | Vals AI |
Agents & tools
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| Terminal-Bench 2.1 (Vals) | – | 76.0%#14 | Vals AI |
| GDPval (AA)not in index | 59.7%#5 | 59.8%#4 | Artificial Analysis |
| Finance Agent v2not in index | – | 49.2%#30 | Vals AI |
| Harvey's Legal Agent Benchmarknot in index | 19.6%#5 | – | Vals AI |
| Legal Research Benchnot in index | 39.9%#19 | – | Vals AI |
| Tax Agent Benchnot in index | – | 62.2%#14 | Vals AI |
| Terminal-Bench 4.0 (Vals)not in index | – | 12.1%#16 | Vals AI |
| Terminal-Bench Sciencenot in index | – | 2.9%#17 | Vals AI |
Knowledge
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| AA-Omniscience | 30.9#18 | 32.0#15 | Artificial Analysis |
| LegalBench | 85.1%#24 | – | Vals AI |
| BioMysteryBenchnot in index | – | 68.5%#6 | Vals AI |
| Excel Modeling Benchmarknot in index | 54.9%#33 | – | Vals AI |
| MedCodenot in index | 48.7%#22 | – | Vals AI |
| MedScribenot in index | 87.2%#12 | – | Vals AI |
| Public Benefits Benchnot in index | 68.5%#4 | – | Vals AI |
Long context
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| AA-LCR | 77.0%#99 | 76.7%#101 | Artificial Analysis |
Composite
| Benchmark | high | xhigh | Source |
|---|---|---|---|
| AA Intelligence Index | 46.3#17 | 46.5#16 | Artificial Analysis |
| Vals Indexnot in index | – | 54.1#23 | Vals AI |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| SpaceXAI | 61 tok/s | 1.38 s | $1.60 | $4.80 | 500k | – |
| SpaceXAI | 55 tok/s | 1.41 s | $1.60 | $4.80 | 500k | – |
| SpaceXAI | 50 tok/s | 1.30 s | $3.20 | $9.60 | 500k | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.500 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0026 | $0.0022 | 6.9 s |
| Summarise a 30-page report | 12,000 / 600 | $0.028 | $0.014 | 12.3 s |
| Code edit | 6,000 / 1,500 | $0.021 | $0.014 | 28.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.144 | $0.076 | 1.2 min |
| Structured extraction | 2,000 / 200 | $0.0052 | $0.0030 | 5.0 s |
See also
Data as of 21 Sept 2026. Compare these configurations.
Cite as: BenchLeader, “Grok 4.7: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/grok-4-7, data as of 21 Sept 2026.