SpaceXAIGrok 4
Grok 4 is a SpaceXAI proprietary reasoning model, released 9 Jul 2025. Its best configuration ranks #161 of 785 on the BenchLeader Index at 58.0 ±2.9, in the upper half. The ± is the point: 230 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in long context (68) and lowest in multimodal (52). At $6.00 per million tokens blended it is among the most expensive ranked models. Output speed of 78 tokens per second puts it slower than most, with a first answer in 7.1 s. It has been measured at 3 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.
- Blended price
- $6.00/M
- $3.00 in · $15.00 out
- Output speed
- 78 tok/s
- measured by Artificial Analysis
- First answer
- 7.07 s
- Context
- 256k
- Full answer
- 14 s
- median, reasoning included
- Cost per run
- –
- Released
- 9 Jul 2025
- Overall index58
- Reasoning55
- Coding61
- Agents & tools58
- Maths58
- Knowledge58
- Instruction following63
- Human preference59
- Multimodal52
- Long context68
- Composite55
Versions
SpaceXAI has shipped 11 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.
| Model | Released | Index | Rank |
|---|---|---|---|
| Grok 4.7high | 21 Sept 2026 | 63.9 | #59 |
| Grok 4.6medium | 11 Aug 2026 | 64.2 | #56 |
| Grok 4.5high | 8 Jul 2026 | 59.3 | #134 |
| Grok 4.3medium | 17 Apr 2026 | 58.7 | #147 |
| Grok 4.2high | 10 Mar 2026 | – | – |
| Grok 4.20thinking | 17 Feb 2026 | 58.6 | #148 |
| Grok 4.1 | 17 Nov 2025 | 57.2 | #184 |
| Grok 4this page | 9 Jul 2025 | 58.0 | #161 |
| Grok 3 | 17 Feb 2025 | 51.5 | #340 |
| Grok 2 | 12 Dec 2024 | 41.9 | #612 |
| Grok 4.20 1 | – | – | – |
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.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| high | – | – | 78 tok/s | 7.07 s | $0.0057 | Agents & tools 34 |
| thinking | – | – | 78 tok/s | 7.07 s | $0.0057 | Reasoning 44 |
| not statedbest | 58.0 | #161 | 78 tok/s | 7.07 s | $0.0057 | Agents & tools 58 · Coding 61 · Composite 55 · Human preference 59 · Instruction following 63 · Knowledge 58 · Long context 68 · Maths 58 · Multimodal 52 · Reasoning 55 |
How its index has moved
8 Sept 2026 to 18 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. 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 | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| GPQA Diamondeffort not stated | – | 87.0%#76 | Epoch AI Benchmarking Hub | |
| SimpleBencheffort not stated | – | 60.5%#30 | SimpleBench | |
| Arena Hard Promptseffort not stated | – | 1421#145 | Arena (LMArena) | |
| GPQA Diamond (AA)not in indexeffort not stated | – | 87.7%#96 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in indexeffort not stated | – | 26.7%#178 | Artificial Analysis | |
| GPQA Diamond (Vals)not in indexeffort not stated | – | 88.1%#34 | Vals AI | |
| Kagi LLM Benchmarkeffort not stated | – | 73.6%#17 | Kagi LLM Benchmark | |
| GPQA Diamond (HELM)not in indexeffort not stated | – | 72.6%#7 | HELM Capabilities | |
| ARC-AGI-1 | 66.7%#151 | 79.6%#121 | ARC Prize | |
| ARC-AGI-2 | 16.0%#150 | 29.4%#133 | ARC Prize | |
| CritPteffort not stated | – | 2.0%#201 | Artificial Analysis | |
| ForecastBenchnot in indexeffort not stated | – | 65.0%#16 | ForecastBench |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| WeirdML | 45.7%#83 | WeirdML | |
| Arena Coding | 1435#163 | Arena (LMArena) | |
| LiveCodeBench | 83.3%#48 | Vals AI | |
| SWE-bench (Vals)not in index | 57.8%#76 | Vals AI | |
| Aider Polyglot | 79.6%#5 | Aider polyglot leaderboard |
Agents & tools
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | 27.2%#62 | Terminal-Bench | |
| GDPval | 21.1%#10 | – | OpenAI | |
| Cybencheffort not stated | – | 43.0%#4 | Cybench | |
| Terminal-Bench Hardeffort not stated | – | 37.9%#62 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in indexeffort not stated | – | 74.8%#139 | Artificial Analysis | |
| BFCL Overalleffort not stated | – | 63.0%#8 | Berkeley Function Calling Leaderboard | |
| Terminal-Bench 2.0 (Vals)not in indexeffort not stated | – | 28.1%#46 | Vals AI | |
| BALROGnot in indexeffort not stated | – | 43.6%#9 | BALROG | |
| DeepResearch Benchnot in indexeffort not stated | – | 47.3%#25 | DeepResearch Bench | |
| METR Time Horizonsnot in indexeffort not stated | – | 66.6%#14 | METR |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| OTIS Mock AIME | 84.0%#104 | Epoch AI Benchmarking Hub | |
| AIME (Vals) | 90.6%#27 | Vals AI | |
| MGSM | 90.9%#30 | Vals AI | |
| IMO 2025 | 21.4%#3 | MathArena | |
| MathArena Apex | 2.1%#29 | MathArena | |
| Omni-MATH (HELM) | 60.3%#9 | HELM Capabilities | |
| MATH 500not in index | 96.2%#2 | Vals AI | |
| Arena Maths | 1425#111 | Arena (LMArena) |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | 2.1#121 | Artificial Analysis | |
| MMLU-Pro | 85.3%#57 | Vals AI | |
| LegalBench | 83.2%#56 | Vals AI | |
| CorpFin | 66.0%#23 | Vals AI | |
| TaxEval | 65.1%#111 | Vals AI | |
| MedQA | 92.5%#31 | Vals AI | |
| MMLU-Pro (HELM)not in index | 85.1%#8 | HELM Capabilities | |
| CaseLaw v2not in index | 65.8%#9 | Vals AI | |
| MedCodenot in index | 38.1%#70 | Vals AI | |
| MedScribenot in index | 78.2%#64 | Vals AI | |
| AA-Omniscience: accuracynot in index | 40.5%#121 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 35.5%#185 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFBench | 53.7%#156 | Artificial Analysis | |
| IFEval (HELM) | 94.9%#2 | HELM Capabilities | |
| WildBench (HELM) | 79.7%#36 | HELM Capabilities | |
| Arena Instruction Followingnot in index | 1396#141 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1411#139 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1394#112 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1413#139 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1416#139 | Arena (LMArena) | |
| Lech Mazur Creative Writingnot in index | 76.9%#21 | Lech Mazur |
Multimodal
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Vision | 1210#86 | Arena (LMArena) | |
| MMMU-Pro | 68.8%#160 | Artificial Analysis | |
| MMMU-Pro (Vals)not in index | 76.3%#52 | Vals AI | |
| MortgageTaxnot in index | 44.5%#83 | Vals AI | |
| SAGEnot in index | 25.1%#85 | Vals AI | |
| GeoBenchnot in index | 45.0%#24 | GeoBench |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Fiction.LiveBench 120k | 96.9%#2 | Fiction.live | |
| AA-LCR | 68.0%#238 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 146.2#75 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 22.5#199 | Artificial Analysis | |
| HELM Capabilities meannot in index | 78.5%#7 | HELM Capabilities |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Price history
The price above, per 1M tokens, every day it changed.
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. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0057 | – | 10.9 s |
| Summarize a 30-page report | 12,000 / 600 | $0.045 | – | 14.8 s |
| Code edit | 6,000 / 1,500 | $0.041 | – | 26.3 s |
| Agentic coding session | 60,000 / 4,000 | $0.240 | – | 58.5 s |
| Structured extraction | 2,000 / 200 | $0.0090 | – | 9.6 s |
More from SpaceXAI
See also
Questions people ask
- What are Grok 4's benchmark scores?
- As of 24 Oct 2026, Grok 4 scores 58.0 ±2.9 on the BenchLeader Index, strongest in long context (68) and weakest in multimodal (52). That rests on 37 independent benchmark results, and every score on this page links to its publisher.
- How does Grok 4 rank?
- Grok 4 is #161 of 785 ranked configurations on 24 Oct 2026. 230 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Grok 4 cost per 1M tokens?
- As of 24 Oct 2026, Grok 4 costs $3.00 per 1M input tokens and $15.00 per 1M output tokens, $6.00 blended (three input tokens per output token). The list price last changed on 9 Sept 2026.
- When was Grok 4 released?
- SpaceXAI released Grok 4 on 9 Jul 2025. BenchLeader has results for it at 3 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Grok 4: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/grok-4.