SpaceXAIGrok 4 Fast
Grok 4 Fast is a SpaceXAI proprietary reasoning model, released 19 Sept 2025. Its best configuration (thinking reasoning effort) ranks #345 of 785 on the BenchLeader Index at 51.4 ±4.1, in the upper half. The ± is the point: 304 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 (62) and lowest in coding (42). At $0.275 per million tokens blended it is cheaper than most ranked models. It has been measured at 4 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 23 Oct 2026.
- Blended price
- $0.275/M
- $0.200 in · $0.500 out
- Output speed
- –
- First answer
- –
- Context
- 2M
- Full answer
- –
- Cost per run
- –
- Released
- 19 Sept 2025
- Overall index51
- Reasoning46
- Coding42
- Agents & tools50
- Maths59
- Knowledge54
- Instruction following52
- Human preference58
- Multimodal44
- Long context62
- Composite50
Versions
SpaceXAI has shipped 2 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 4.1 Fastthinking | 19 Nov 2025 | 54.7 | #244 |
| Grok 4 Fastthinkingthis page | 19 Sept 2025 | 51.4 | #345 |
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 |
|---|---|---|---|---|---|---|
| no reasoning | 44.2 | #543 | – | – | $0.0002 | Agents & tools 44 · Coding 40 · Composite 42 · Human preference 60 · Instruction following 41 · Knowledge 42 · Long context 36 · Maths 51 · Multimodal 30 · Reasoning 47 |
| high | – | – | – | – | – | |
| thinkingbest | 51.4 | #345 | – | – | $0.0002 | Agents & tools 50 · Coding 42 · Composite 50 · Human preference 58 · Instruction following 52 · Knowledge 54 · Long context 62 · Maths 59 · Multimodal 44 · Reasoning 46 |
| not stated | – | – | – | – | – | Agents & tools 51 · Coding 45 · Long context 63 · Reasoning 34 |
How its index has moved
8 Sept 2026 to 21 Oct 2026- thinking
- no reasoning
The 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 | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Arena Hard Prompts | 1432#136 | 1414#158 | – | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 60.6%#371 | 84.8%#133 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 4.5%#493 | 19.1%#223 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | 62.1%#104 | 85.3%#47 | – | Vals AI | |
| Kagi LLM Benchmark | – | 66.1%#38 | 35.4%#116 | Kagi LLM Benchmark | |
| ARC-AGI-1 | – | 48.5%#181 | – | ARC Prize | |
| ARC-AGI-2 | – | 5.3%#174 | – | ARC Prize | |
| CritPt | 0.0%#315 | 2.9%#182 | – | Artificial Analysis | |
| DTBenchnot in indexeffort not stated | – | – | 82.7%#117 | Epoch AI Benchmarking Hub | |
| ForecastBenchnot in indexeffort not stated | – | – | 64.9%#17 | ForecastBench |
Coding
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| WeirdMLeffort not stated | – | – | 42.9%#91 | WeirdML | |
| Arena Coding | 1457#137 | 1439#157 | – | Arena (LMArena) | |
| Arena WebDev | – | 1160#138 | – | Arena (LMArena) | |
| LiveCodeBench | 46.1%#115 | 79.0%#73 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | 45.4%#81 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | 0.0%#112 | – | Vals AI |
Agents & tools
| Benchmark | no reasoning | high | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| Cybencheffort not stated | – | – | – | 30.0%#9 | Cybench | |
| Terminal-Bench Hard | 12.1%#210 | – | 18.9%#166 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 63.7%#171 | – | 65.8%#165 | – | Artificial Analysis | |
| τ³-bench Bankingnot in index | – | 15.7%#21 | – | – | τ²-bench | |
| Poker Agentnot in index | – | – | 1034.3%#9 | – | Vals AI | |
| Terminal-Bench 2.0 (Vals)not in index | – | – | 29.2%#45 | – | Vals AI |
Maths
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AIME (Vals) | 33.3%#69 | 91.3%#25 | Vals AI | |
| MGSM | 88.0%#51 | 90.9%#31 | Vals AI | |
| Arena Maths | 1424#113 | 1412#131 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-Omniscience | -54.4#428 | -29.9#282 | Artificial Analysis | |
| MMLU-Pro | 70.3%#112 | 79.7%#89 | Vals AI | |
| LegalBench | 78.4%#98 | 80.6%#83 | Vals AI | |
| CorpFin | 58.4%#81 | 66.9%#15 | Vals AI | |
| TaxEval | 71.6%#74 | 75.7%#13 | Vals AI | |
| MedQA | 75.4%#76 | 92.1%#35 | Vals AI | |
| CaseLaw v2not in index | – | 65.7%#10 | Vals AI | |
| MedCodenot in index | 30.0%#93 | 37.4%#73 | Vals AI | |
| MedScribenot in index | 79.7%#59 | 81.6%#51 | Vals AI | |
| AA-Omniscience: accuracynot in index | 17.7%#357 | 22.8%#268 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 12.3%#376 | 31.7%#211 | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| IFBench | 37.7%#290 | 50.5%#172 | Artificial Analysis | |
| Arena Instruction Followingnot in index | 1395#144 | 1389#150 | Arena (LMArena) |
Human preference
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| Arena Text | 1419#125 | 1405#144 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1382#134 | 1384#130 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1422#121 | 1405#146 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1418#136 | 1413#141 | Arena (LMArena) |
Multimodal
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| MMMU-Pro | 48.1%#247 | 61.8%#205 | Artificial Analysis | |
| MMMU-Pro (Vals)not in index | 63.4%#78 | 72.8%#59 | Vals AI | |
| MortgageTaxnot in index | 25.3%#91 | 42.1%#86 | Vals AI | |
| SAGEnot in index | 16.0%#88 | 29.8%#81 | Vals AI |
Long context
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Fiction.LiveBench 120keffort not stated | – | – | 75.0%#7 | Fiction.live | |
| AA-LCR | 24.0%#415 | 73.7%#173 | – | Artificial Analysis |
Composite
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 144.1#89 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 11.1#365 | 17.9#258 | – | Artificial Analysis |
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 | $0.0002 | – | – |
| Summarize a 30-page report | 12,000 / 600 | $0.0027 | – | – |
| Code edit | 6,000 / 1,500 | $0.0020 | – | – |
| Agentic coding session | 60,000 / 4,000 | $0.014 | – | – |
| Structured extraction | 2,000 / 200 | $0.0005 | – | – |
More from SpaceXAI
See also
Questions people ask
- What are Grok 4 Fast's benchmark scores?
- As of 25 Oct 2026, Grok 4 Fast (thinking effort) scores 51.4 ±4.1 on the BenchLeader Index, strongest in long context (62) and weakest in coding (42). That rests on 23 independent benchmark results, and every score on this page links to its publisher.
- How does Grok 4 Fast rank?
- Grok 4 Fast (thinking effort) is #345 of 785 ranked configurations on 25 Oct 2026. 304 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 Fast cost per 1M tokens?
- As of 25 Oct 2026, Grok 4 Fast costs $0.200 per 1M input tokens and $0.500 per 1M output tokens, $0.275 blended (three input tokens per output token). The list price last changed on 9 Sept 2026.
- When was Grok 4 Fast released?
- SpaceXAI released Grok 4 Fast on 19 Sept 2025. BenchLeader has results for it at 4 reasoning-effort settings.
Data as of 25 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Grok 4 Fast: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/grok-4-fast.