Grok 4.7 vs Muse Spark Open
Full profiles: Grok 4.7 (SpaceXAI) and Muse Spark Open (Meta): every benchmark, price and rank.
Updated · Data from 14 sources
As of 16 Nov 2026, Grok 4.7 and Muse Spark Open are neck and neck on quality (57.4 vs 53.5 on the BenchLeader Index, inside the ±7.0 margin), and Muse Spark Open is 1.5× cheaper at $2.00 vs $3.00 per 1M tokens blended. Choose Grok 4.7 for coding (61 vs 36), reasoning (56 vs 51), and knowledge (60 vs 59); choose Muse Spark Open for fast first answers (9.9 s vs 55.9 s) and multimodal (63 vs 56), where it also streams ~2× faster.
Key facts (as of 16 Nov 2026)
- Grok 4.7 scores 57.4 on the BenchLeader Index and Muse Spark Open scores 53.5 (BenchLeader, 16 Nov 2026).
- Grok 4.7 costs $3.00 per 1M tokens blended (list price, 16 Nov 2026).
- Muse Spark Open costs $2.00 per 1M tokens blended (list price, 16 Nov 2026).
- Muse Spark Open streams ~2× faster than Grok 4.7 (Artificial Analysis, 7-day median, 16 Nov 2026).
- Of the 39 benchmarks both models have, Grok 4.7 leads 31 (BenchLeader, 16 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Grok 4.7 | 61 vs 36 category score; FrontierCode 2.0 (AA) 28.7% vs 7.6% | High |
| Reasoning | Grok 4.7 | 56 vs 51 category score; ARC-AGI-2 61.4% vs 57.5% | Medium |
| Knowledge | Grok 4.7 | 60 vs 59 category score; AA-Omniscience v1.1 31.2 vs 25.8 | Medium |
| Long documents | Either | 63 vs 64 category score; AA-LCR v1.1 76.7% vs 78.3% | Either |
| Images and other media | Muse Spark Open | 63 vs 56 category score; SAGE 85.8% vs 88.7% | High |
| Chat quality (human preference) | – | Only Grok 4.7 has an Arena rating (63 category score) | Not comparable |
| Lowest cost | Muse Spark Open | $2.00 vs $3.00 per 1M blended; agentic coding session $0.092 vs $0.144 | – |
| Fastest first answer | Muse Spark Open | 9.9 s (Muse Spark Open (xhigh)) vs 55.9 s (Grok 4.7 (xhigh)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Grok 4.7 (xhigh) and Muse Spark Open (xhigh) are neck and neck on quality (57.4 vs 53.5, within the ±7.0 margin).
- Grok 4.7 (xhigh) is stronger in coding, knowledge, reasoning.
- Muse Spark Open (xhigh) is stronger in long context, multimodal.
- Muse Spark Open (xhigh) is 1.5× cheaper ($2.00 vs $3.00 per 1M blended).
- Muse Spark Open (xhigh) streams 2.7× faster (180 vs 68 tokens per second).
At a glance
| Spec | Grok 4.7 | Muse Spark Open |
|---|---|---|
| Developer | SpaceXAI | Meta |
| Released | 21 Sep 2026 | 15 Oct 2026 |
| Context window (tokens) | 500,000 | 1,048,576 |
| Compared at | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | $2.00 / $6.00 / $0.500 | $1.25 / $4.25 / $0.150 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 55.6 | 51.3 | Grok 4.7 |
| Coding | 60.9 | 36.0 | Grok 4.7 |
| Agents & tools | 56.9 | – | not comparable |
| Maths | 55.8 | – | not comparable |
| Knowledge | 59.9 | 58.8 | Grok 4.7 |
| Human preference | 62.5 | – | not comparable |
| Multimodal | 55.5 | 63.3 | Muse Spark Open |
| Long context | 62.9 | 63.8 | level |
| Composite | 52.8 | – | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Agents & tools, Maths, Instruction following, Human preference.
Where sources disagree
Across the 39 benchmarks both have, Grok 4.7 leads 31, Muse Spark Open leads 8. Every overall index that rates both puts Grok 4.7 ahead.
| Overall index | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) | Leader |
|---|---|---|---|
| BenchLeader Index | 57.4 | 53.5 | neck and neck |
| AA Intelligence Index v5.0.1 | 35.5 | 24.9 | Grok 4.7 |
| Vals Index | 58.6 | 48.6 | Grok 4.7 |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $3.00 | $2.00 |
| Input / output per 1M | $2.00 / $6.00 | $1.25 / $4.25 |
| Cached input per 1M | $0.500 | $0.150 |
| Price per task: chat reply | $0.0026 | $0.0018 |
| Price per task: summarize a 30-page report | $0.028 | $0.018 |
| Price per task: code edit | $0.021 | $0.014 |
| Price per task: agentic coding session | $0.144 | $0.092 |
| Price per task: structured extraction | $0.0052 | $0.0034 |
| Agentic coding session, input mostly cached | $0.076 | $0.043 |
| Cost per benchmark task (AA Index) | $7.77 | $2.02 |
| Output tokens per task (AA Index) | 105k | 66k |
| Of which reasoning | 69k | 32k |
| Time per task (AA Index) | 26 min | 6 min |
| Output speed | 68 tok/s | 180 tok/s |
| Time to first answer | 55.9 s | 9.9 s |
| Time to a complete answer, last measured 13 Nov 2026 | 145.5 s | 39.9 s |
Per token is not per task: an average task of Artificial Analysis' Intelligence Index cost $7.77 on Grok 4.7 (xhigh) and $2.02 on Muse Spark Open (xhigh), and Grok 4.7 (xhigh) wrote 1.6× as many output tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. Grok 4.7 (xhigh) scores 57.4 and takes 56 s to its first answer token; Grok 4.7 (high) scores 56.9 in 35 s, Grok 4.7 (low) scores 55.4 in 10 s. Muse Spark Open (xhigh) scores 53.5 and takes 10 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
159 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 57.4 | 53.5 |
| Agents & tools score | BenchLeader | 56.9 | – |
| Coding score | BenchLeader | 60.9 | 36.0 |
| Composite score | BenchLeader | 52.8 | – |
| Human preference score | BenchLeader | 62.5 | – |
| Knowledge score | BenchLeader | 59.9 | 58.8 |
| Long context score | BenchLeader | 62.9 | 63.8 |
| Maths score | BenchLeader | 55.8 | – |
| Multimodal score | BenchLeader | 55.5 | 63.3 |
| Reasoning score | BenchLeader | 55.6 | 51.3 |
| Blended price $/M | BenchLeader | $3.00 | $2.00 |
| Output speed | BenchLeader | 68 tok/s | 180 tok/s |
| Time to first answer | BenchLeader | 55.9 s | 9.9 s |
| Context window | BenchLeader | 500,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 92.7% | – |
| Arena Hard Prompts | Arena (LMArena) | 1466 | – |
| LiveBench 2.0 Reasoning | LiveBench | 82.7% | 73.7%† |
| LiveBench 1.0 Reasoning | LiveBench | 82.7% | – |
| GPQA Diamond (AA) | Artificial Analysis | – | 89.0% |
| Humanity's Last Exam (AA) | Artificial Analysis | 43.1% | 43.9% |
| HLE-Diamond (AA) | Artificial Analysis | 23.2% | 26.3% |
| ARC-AGI-1 | ARC Prize | 89.5% | 89.0% |
| ARC-AGI-2 | ARC Prize | 61.4% | 57.5% |
| ARC-AGI-3 | ARC Prize | 7.0% | 0.4% |
| CritPt | Artificial Analysis | 17.7% | 14.9% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 33.5% | 29.5% |
| MysteryMechanism | Vals AI | 25.2% | – |
| HLE-Diamond (Scale) | Scale AI SEAL | 23.4%† | 28.0%† |
| Arena Expert | Arena (LMArena) | 1478 | – |
| Chess Puzzles | Epoch AI Benchmarking Hub | 41.0% | – |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 29.0% | – |
| LMCA | Epoch AI Benchmarking Hub | 49.4% | – |
| DTBench | Epoch AI Benchmarking Hub | 96.0% | – |
| Coding | |||
| SciCode | SciCode | 57.4% | – |
| FrontierCode | Cognition | 47.6%† | – |
| Arena Coding | Arena (LMArena) | 1493 | – |
| Arena WebDev | Arena (LMArena) | 1641 | 1526† |
| LiveBench 2.0 Coding | LiveBench | 73.1% | 55.1%† |
| LiveBench 1.0 Coding | LiveBench | 77.2% | – |
Show all 159 benchmarks
| Metric | Publisher | Grok 4.7 (xhigh) | Muse Spark Open (xhigh) |
|---|---|---|---|
| Coding | |||
| SciCode (AA) | Artificial Analysis | 57.4% | 58.5% |
| IOI | Vals AI | 57.7% | – |
| Code Migration | Vals AI | 44.8% | 21.2% |
| Code Migration v1.1 | Vals AI | 53.2% | 26.9% |
| ProgramBench | Vals AI | 0.5% | – |
| SWE-bench Multimodal | Vals AI | 42.9% | – |
| Vibe Code Bench v1.1 | Vals AI | 86.2% | 76.2% |
| Vibe Code Bench 1-100 v1.1 | Vals AI | 26.4% | – |
| SWE Atlas v1.1 | Scale AI SEAL | 62.1%† | 42.3%† |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 73.4%† | 50.0%† |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 64.4%† | 41.1%† |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 48.6%† | 35.7%† |
| SWE-Bench Pro V2 | Scale AI SEAL | – | 95.0% |
| SWE-Bench Pro V2 HARD | Scale AI SEAL | – | 82.3% |
| MirrorCode | Epoch AI Benchmarking Hub | 54.4% | – |
| CursorBench | Cursor | 46.3% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 1545.8 | – |
| FrontierSWE | Epoch AI Benchmarking Hub | 29.5% | – |
| WeirdML v3 | WeirdML | 8.4% | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 25.8% | 8.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 76.4% |
| Startrise Frontend | Startrise | 75.9%† | – |
| GameBench 2 Overall | DuelLab GameBench 2 | 63.6† | 55.3† |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 71.2 | 59.3 |
| Design Arena Website | Design Arena | 1244† | 1305 |
| FrontierCode 1.1 (AA) | Artificial Analysis | 33.8% | 27.6% |
| FrontierCode 2.0 (AA) | Artificial Analysis | 28.7% | 7.6% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 38.8% | – |
| APEX-Agents | Mercor | 54.6%† | – |
| Arena Agent | Arena (LMArena) | 2.6 | – |
| LiveBench Long Agentic | LiveBench | 45.3% | 12.1%† |
| LiveBench Browser | LiveBench | 49.3% | 26.2%† |
| LiveBench Agentic Coding | LiveBench | 54.0% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 73.4% | – |
| HiL-Bench | Scale AI SEAL | – | 59.7%† |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 60.7% | 45.1% |
| GDPval-AA v2.1 | Artificial Analysis | 1717 | 1401 |
| ITBench SRE (AA) | Artificial Analysis | 42.1% | – |
| Analyst Agent (AA) | Artificial Analysis | 52.5% | – |
| CyberBench | Vals AI | 69.5% | – |
| Finance Agent v2 | Vals AI | 52.3% | 61.3% |
| Finance Agent v2.1 | Vals AI | 60.6% | 60.7% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 18.3% | 25.0% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 12.5% | 16.7% |
| Legal Research Bench | Vals AI | 47.1% | 38.5% |
| Robo-bench | Vals AI | 0.2% | – |
| SREBench | Vals AI | 3.8% | – |
| Tax Agent Bench | Vals AI | 35.8% | 26.3% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 28.8% | 6.6% |
| Terminal-Bench Science | Vals AI | 10.0% | – |
| Arena Agent Task Completion | Arena (LMArena) | 5.1 | – |
| Vending-Bench 2 | Andon Labs | 10536.8† | – |
| AutomationBench | Artificial Analysis | 65.6% | 49.7% |
| GDP.pdf (AA) | Artificial Analysis | 20.0% | 18.6% |
| Harvey LAB-AA v1.1 | Harvey | 9.4% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1636 | 1280 |
| GDPx | GDPx by SemiAnalysis | 3.9% | 1.7%† |
| Design Arena Agentic | Design Arena | 1277† | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 11.4% | 1.9% |
| CWE-Bench (AA) | Artificial Analysis | 68.3% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 26.9% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 74.0% | – |
| MKA (AA) | Artificial Analysis | 40.1% | – |
| FACTS Search | 85.6%† | – | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 53.0% | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 17.1% | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 98.1% | – |
| ProofBench | Vals AI | 26.0%† | – |
| LiveBench Mathematics | LiveBench | 95.7% | – |
| ArXivMath | MathArena | – | 27.0% |
| BrokenArXiv | MathArena | – | 30.1% |
| Arena Maths | Arena (LMArena) | 1447 | – |
| FrontierMath Erdős | Epoch AI Benchmarking Hub | 0.0% | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 56.0% | – |
| LiveBench 2.0 Data Analysis | LiveBench | 76.9% | 60.2%† |
| LiveBench Research | LiveBench | 89.5% | 82.7%† |
| LiveBench 1.0 Data Analysis | LiveBench | 76.9% | – |
| AA-Omniscience v1.1 | Artificial Analysis | 31.2 | 25.8 |
| LegalBench | Vals AI | 84.4% | – |
| PRBench Legal | Scale AI SEAL | 47.6% | 59.8%† |
| BioMysteryBench | Vals AI | 69.3% | – |
| Excel Modeling Benchmark | Vals AI | 67.0% | 53.7% |
| Excel Modeling Benchmark v1.1 | Vals AI | 71.5% | 62.0% |
| MedCode | Vals AI | 49.5% | – |
| MedScribe | Vals AI | 89.4% | – |
| Public Benefits Bench v1.1 | Vals AI | 65.6% | – |
| Access to Justice Benchmark | Vals AI | 73.7% | – |
| Financial Coaching Bench | Vals AI | 74.1% | – |
| Immigration Bench | Vals AI | 68.0% | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 22.8% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 42.7% | 42.0% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 80.0% | 72.0% |
| HealthBench Professional (AA) | OpenAI | 55.8% | – |
| FACTS Parametric | 54.3%† | – | |
| Instruction following | |||
| LiveBench Language | LiveBench | 80.1% | – |
| LiveBench Instruction Following | LiveBench | 75.3% | – |
| MultiChallenge | Scale AI SEAL | – | 74.2%† |
| Arena Instruction Following | Arena (LMArena) | 1435 | – |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1443 | – |
| EQ-Bench Creative Writing v3 | EQ-Bench | 2012† | – |
| Arena Creative Writing | Arena (LMArena) | 1421 | – |
| Arena Multi-turn | Arena (LMArena) | 1435 | – |
| Arena Longer Queries | Arena (LMArena) | 1454 | – |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1230 | – |
| LiveBench Design | LiveBench | 75.0% | 58.1%† |
| MMMU-Pro | Artificial Analysis | 78.2%† | 80.9% |
| SAGE | Vals AI | 88.7% | 85.8% |
| Humanity's Sixth Sense | Scale AI SEAL | – | 29.6%† |
| GeoBench | GeoBench | 61.0%† | – |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 32.5%† | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 20.8% | – |
| VIALS-AA | Handshake AI | 37.9% | – |
| Code2Video-AA | Artificial Analysis | 1127 | – |
| AA-3D | Artificial Analysis | 992 | – |
| FACTS Multimodal | 40.9%† | – | |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 76.7% | 78.3% |
| MLCR | Artificial Analysis | 20.0% | 7.8% |
| FACTS Grounding | 68.6%† | – | |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 156.2† | 155.4† |
| LiveBench 2.0 | LiveBench | 67.9% | 48.5%† |
| LiveBench 1.0 | LiveBench | 77.4% | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 35.5 | 24.9 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 46.5 | 38.9 |
| AA Progress Index v1 | Artificial Analysis | 322.7 | 306.7 |
| Vals Index | Vals AI | 58.6 | 48.6 |
| Vals Finance Index | Vals AI | 58.8% | 53.0% |
| Vals RSI Index | Vals AI | 20.2% | – |
| Vals Science Index | Vals AI | 34.8% | – |
| Vals Router Index | Vals AI | 58.6%† | – |
| Vals Legal Index | Vals AI | 72.7% | 68.5% |
| AA Cyber Index | Artificial Analysis | 56.4 | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | – | 7.6%† |
| DistressBench | Scale AI SEAL | 52.4%† | 87.7%† |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 27.8 |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- Grok 4.7 (xhigh): BenchLeader Index 58.0 → 57.4.
- Grok 4.7 (xhigh): 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
- Muse Spark Open (xhigh): 1 score revision.
- Grok 4.7 (xhigh): 1 new result (MirrorCode).
- Grok 4.7 (xhigh): 1 new result (ALE-Bench).
- Grok 4.7 (xhigh): 2 score revisions.
- Grok 4.7 (xhigh): 2 score revisions.
- Muse Spark Open (xhigh): 1 new result (GameBench 2 by reasoning setting); 1 score revision.
- Grok 4.7 (xhigh): 2 score revisions.
- Muse Spark Open (xhigh): BenchLeader Index 53.3 → 53.8.
Related comparisons
- Grok 4.7 vs Claude Opus 5.5
- Grok 4.7 vs GPT-6 Astra
- Grok 4.7 vs Claude Fable 5.1
- Muse Spark Open vs Claude Opus 5.5
- Muse Spark Open vs GPT-6 Astra
- Muse Spark Open vs GPT-6.1 Sol
- Muse Spark Open vs Claude Fable 5
- Grok 4.7 vs Muse Wonder
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Grok 4.7 vs Muse Spark Open: questions
- Is Grok 4.7 better than Muse Spark Open?
- Grok 4.7 (xhigh) and Muse Spark Open (xhigh) are neck and neck on quality (57.4 vs 53.5, within the ±7.0 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
- Grok 4.7 vs Muse Spark Open: which is cheaper, and will that change?
- Muse Spark Open is cheaper today: $2.00 against $3.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Grok 4.7 or Muse Spark Open better for coding?
- Grok 4.7 scores higher in coding (61 vs 36 on the category index, where 50 is average). On FrontierCode 2.0 (AA), the top shared coding benchmark, it is 28.7% for Grok 4.7 against 7.6% for Muse Spark Open.
- Which is faster, Grok 4.7 or Muse Spark Open?
- Muse Spark Open streams ~2× faster (180 vs 68 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- They are level on long-context benchmarks (63 vs 64). Muse Spark Open accepts more context: 1,048,576 against 500,000 tokens for Grok 4.7.
More questions (3)
- Which has lower latency?
- Muse Spark Open answers first: 9.9 s against 55.9 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on Grok 4.7 vs Muse Spark Open?
- Across the 39 benchmarks both have, Grok 4.7 leads 31, Muse Spark Open leads 8. Every overall index that rates both puts Grok 4.7 ahead.
- When was Muse Spark Open released?
- Muse Spark Open was released on 15 Oct 2026.