GPT-6 Astra vs Muse Spark 1.3
Full profiles: GPT-6 Astra (OpenAI) and Muse Spark 1.3 (Meta): every benchmark, price and rank.
Updated · Data from 10 sources
As of 24 Oct 2026, GPT-6 Astra (max) and Muse Spark 1.3 (max) are neck and neck on quality (70.4 vs 67.1 on the BenchLeader Index, inside the ±5.8 margin of error), and Muse Spark 1.3 (max) costs 10× less, which makes it the better buy for most workloads. Pick GPT-6 Astra for coding (74 vs 65); pick Muse Spark 1.3 for human preference (69 vs 67).
GPT-6 Astra benchmarks 
Muse Spark 1.3 benchmarks 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 | GPT-6 Astra | 74 vs 65 category score; Arena WebDev 1775 vs 1657 | High |
| Agents & tool use | Muse Spark 1.3 | 66 vs 64 category score; τ³-Banking (AA) 50.5% vs 41.4% | Medium |
| Maths | GPT-6 Astra | 72 vs 63 category score; FrontierMath Tiers 1–3 93.7% vs 74.0% | High |
| Reasoning | Muse Spark 1.3 | 77 vs 76 category score; Arena Hard Prompts 1517 vs 1501 | Medium |
| Knowledge | GPT-6 Astra | 80 vs 73 category score; AA-Omniscience 43.4 vs 25 | Medium |
| Long documents | Muse Spark 1.3 | 67 vs 65 category score; AA-LCR 83.0% vs 80.7% | Medium |
| Images and other media | Either | 66 vs 66 category score; Arena Vision 1280 vs 1308 | Either |
| Chat quality (human preference) | Muse Spark 1.3 | 69 vs 67 on Arena votes | Medium |
| Lowest cost | Muse Spark 1.3 | $2.00 vs $20.00 per 1M blended; agentic coding session $0.092 vs $0.800 | – |
| Fastest first answer | Muse Spark 1.3 | 90.1 s (Muse Spark 1.3 (max)) vs 247.0 s (GPT-6 Astra (max)) 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
- GPT-6 Astra (max) and Muse Spark 1.3 (max) are neck and neck on quality (70.4 vs 67.1, within the ±5.8 margin).
- GPT-6 Astra (max) is stronger in coding, composite, knowledge, maths, multimodal.
- Muse Spark 1.3 (max) is stronger in agents & tools, human preference, long context, reasoning.
- Muse Spark 1.3 (max) is 10× cheaper ($2.00 vs $20.00 per 1M blended).
- Muse Spark 1.3 (max) streams 2.0× faster (116 vs 58 tokens per second).
At a glance
| Spec | GPT-6 Astra | Muse Spark 1.3 |
|---|---|---|
| Developer | OpenAI | Meta |
| Released | 2 Sept 2026 | 2 Sept 2026 |
| Context window (tokens) | 1,050,000 | 1,048,576 |
| Compared at | GPT-6 Astra (max) | Muse Spark 1.3 (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $10.00 / $50.00 / $1.00 | $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 | GPT-6 Astra (max) | Muse Spark 1.3 (max) | Leader |
|---|---|---|---|
| Reasoning | 76.2 | 77.4 | Muse Spark 1.3 |
| Coding | 74.0 | 64.5 | GPT-6 Astra |
| Agents & tools | 64.2 | 65.6 | Muse Spark 1.3 |
| Maths | 71.8 | 63.4 | GPT-6 Astra |
| Knowledge | 79.5 | 72.8 | GPT-6 Astra |
| Human preference | 66.7 | 69.1 | Muse Spark 1.3 |
| Multimodal | 66.0 | 65.7 | level |
| Long context | 65.4 | 66.6 | Muse Spark 1.3 |
| Composite | 88.4 | 83.3 | GPT-6 Astra |
Where sources disagree
Across the 56 benchmarks both have, GPT-6 Astra leads 38, Muse Spark 1.3 leads 18. Independent sources split: GPT-6 Astra leads on AA Intelligence Index v4.3.2; Muse Spark 1.3 leads on Arena Text.
| Overall index | GPT-6 Astra (max) | Muse Spark 1.3 (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 70.4 | 67.1 | neck and neck |
| AA Intelligence Index v4.3.2 | 52.7 | 48.1 | GPT-6 Astra |
| Arena Text | 1477 | 1496 | Muse Spark 1.3 |
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; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | GPT-6 Astra (max) | Muse Spark 1.3 (max) |
|---|---|---|
| Blended price per 1M tokens | $20.00 | $2.00 |
| Input / output per 1M | $10.00 / $50.00 | $1.25 / $4.25 |
| Cached input per 1M | $1.00 | $0.150 |
| Price per task: chat reply | $0.019 | $0.0018 |
| Price per task: summarize a 30-page report | $0.150 | $0.018 |
| Price per task: code edit | $0.135 | $0.014 |
| Price per task: agentic coding session | $0.800 | $0.092 |
| Price per task: structured extraction | $0.030 | $0.0034 |
| Cost to run a full benchmark suite | $3.26 | $1.60 |
| Output speed | 58 tok/s | 116 tok/s |
| Time to first answer | 247.0 s | 90.1 s |
| Time to a complete answer | 255.6 s | 94.4 s |
Per token is not per task: on Artificial Analysis' full suite GPT-6 Astra (max) cost $3.26 and Muse Spark 1.3 (max) $1.60, because Muse Spark 1.3 (max) used about 4.9× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. GPT-6 Astra (max) scores 70.4 and takes 247 s to its first answer token; GPT-6 Astra (high) scores 69.6 in 35 s, GPT-6 Astra (xhigh) scores 68.9 in 130 s. Muse Spark 1.3 (max) scores 67.1 and takes 90 s to its first answer token; Muse Spark 1.3 (xhigh) scores 64.9 in 30 s, Muse Spark 1.3 scores 59.8 in 3 s. 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
131 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 | GPT-6 Astra (max) | Muse Spark 1.3 (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 70.4 | 67.1 |
| Agents & tools score | BenchLeader | 64.2 | 65.6 |
| Coding score | BenchLeader | 74.0 | 64.5 |
| Composite score | BenchLeader | 88.4 | 83.3 |
| Human preference score | BenchLeader | 66.7 | 69.1 |
| Knowledge score | BenchLeader | 79.5 | 72.8 |
| Long context score | BenchLeader | 65.4 | 66.6 |
| Maths score | BenchLeader | 71.8 | 63.4 |
| Multimodal score | BenchLeader | 66.0 | 65.7 |
| Reasoning score | BenchLeader | 76.2 | 77.4 |
| Blended price $/M | BenchLeader | $20.00 | $2.00 |
| Output speed | BenchLeader | 58 tok/s | 116 tok/s |
| Time to first answer | BenchLeader | 247.0 s | 90.1 s |
| Context window | BenchLeader | 1,050,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 95.8% | – |
| Humanity's Last Exam | Scale AI / CAIS | 54.8%† | – |
| Arena Hard Prompts | Arena (LMArena) | 1501 | 1517 |
| LiveBench 2.0 Reasoning | LiveBench | 92.7%† | 89.7%† |
| LiveBench 1.0 Reasoning | LiveBench | 92.7% | 89.7%† |
| GPQA Diamond (AA) | Artificial Analysis | 96.1% | 93.5% |
| Humanity's Last Exam (AA) | Artificial Analysis | 54.7% | 48.7% |
| ARC-AGI-1 | ARC Prize | 97.5% | – |
| ARC-AGI-2 | ARC Prize | 95.0% | – |
| ARC-AGI-3 | ARC Prize | 98.6% | – |
| CritPt | Artificial Analysis | 31.7% | 24.9% |
| MysteryMechanism | Vals AI | 53.1% | 36.0% |
| Chess Puzzles | Epoch AI Benchmarking Hub | 74.0% | 37.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 84.0% | 25.0% |
| LMCA | Epoch AI Benchmarking Hub | 64.1%† | 53.9%† |
| DTBench | Epoch AI Benchmarking Hub | 97.1%† | 96.5%† |
| BTF-3 | Epoch AI Benchmarking Hub | 13.5%† | 14.1%† |
| Coding | |||
| SciCode | SciCode | 56.5% | 58.8% |
| WeirdML | WeirdML | 93.3% | – |
| FrontierCode | Cognition | 53.3% | – |
| GSO-Bench | GSO-Bench | 79.4%† | – |
| Arena Coding | Arena (LMArena) | 1543 | 1536 |
| Arena WebDev | Arena (LMArena) | 1775 | 1657 |
| LiveBench 2.0 Coding | LiveBench | 93.7%† | 90.2%† |
| LiveBench 1.0 Coding | LiveBench | 80.4% | 81.1%† |
Show all 131 benchmarks
| Metric | Publisher | GPT-6 Astra (max) | Muse Spark 1.3 (max) |
|---|---|---|---|
| Coding | |||
| IOI | Vals AI | 100.0% | 56.6% |
| Code Migration | Vals AI | 67.7% | 47.4% |
| Code Migration v1.1 | Vals AI | 80.6% | – |
| ProgramBench | Vals AI | 5.5% | 2.5% |
| SWE-bench Multimodal | Vals AI | 43.3% | 43.1% |
| Vibe Code Bench 1-100 | Vals AI | 27.6% | 20.5% |
| Vibe Code Bench v1.1 | Vals AI | 89.6% | 85.9% |
| SWE Atlas: Codebase QnA | Scale AI SEAL | 59.1%† | – |
| SWE Atlas: Refactoring | Scale AI SEAL | 59.0%† | – |
| SWE Atlas: Test Writing | Scale AI SEAL | 50.7%† | – |
| MirrorCode | Epoch AI Benchmarking Hub | 46.7%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 73.2% | – |
| CursorBench | Cursor | – | 41.6% |
| ALE-Bench | Epoch AI Benchmarking Hub | 2951.3 | 1688.3† |
| FrontierSWE | Epoch AI Benchmarking Hub | 65.5% | 25.8% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 59.1% | 33.3% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 88.4% | 84.3% |
| Startrise Frontend | Startrise | 79.9%† | 74.7%† |
| FrontierCode 1.1 (AA) | Artificial Analysis | 47.0% | 40.6% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 58.2% | 14.6%† |
| Remote Labor Index | Scale AI / CAIS | 20.8%† | – |
| APEX-Agents | Mercor | 64.7%† | 47.6% |
| Arena Agent | Arena (LMArena) | 14.7 | 3.4 |
| LiveBench Long Agentic | LiveBench | 82.4%† | 40.4%† |
| LiveBench Browser | LiveBench | 84.0%† | 80.0%† |
| LiveBench Agentic Coding | LiveBench | 57.3% | 64.1%† |
| Terminal-Bench 2.1 (Vals) | Vals AI | 87.3% | 79.0% |
| HiL-Bench | Scale AI SEAL | 65.3%† | 66.9%† |
| GDPval-AA v2.1 | Artificial Analysis | 53.7% | – |
| τ³-Banking (AA) | Artificial Analysis | 41.4% | 50.5% |
| ITBench SRE (AA) | Artificial Analysis | 48.6% | 33.2% |
| Analyst Agent (AA) | Artificial Analysis | 51.3% | 37.5% |
| CUA-bench | Vals AI | 19.2% | 5.8% |
| CyberBench | Vals AI | 41.1% | 72.7% |
| Finance Agent v2 | Vals AI | 53.5% | 60.0% |
| Finance Agent v2.1 | Vals AI | 60.4% | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 7.1% | – |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 5.4% | 23.8% |
| Legal Research Bench | Vals AI | 39.4% | 55.3% |
| Robo-bench | Vals AI | 23.3% | – |
| SREBench | Vals AI | 56.9% | 3.8% |
| Tax Agent Bench | Vals AI | 20.7% | 42.4% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 59.6% | 24.8% |
| Terminal-Bench Science | Vals AI | 62.9% | 10.0% |
| Time Horizon Index: KSP | Vals AI | 90.5% | – |
| Arena Search | Arena (LMArena) | 1246 | – |
| BALROG | BALROG | 68.3% | – |
| Vending-Bench 2 | Andon Labs | 15514.7† | – |
| AutomationBench | Artificial Analysis | 68.5% | 57.9% |
| GDP.pdf (AA) | Artificial Analysis | 31.0% | 26.6% |
| Harvey LAB-AA v1.1 | Harvey | 8.6% | 8.9% |
| AA-Briefcase v1.1 | Artificial Analysis | 1568 | 1581 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 63.3% | 10.9% |
| CWE-Bench (AA) | Artificial Analysis | 63.3% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 36.9% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 0.0% | – |
| FACTS Search | 89.9%† | – | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 93.7% | 74.0% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 97.6% | 46.3% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 100.0% | 99.2%† |
| ProofBench | Vals AI | 99.0%† | 58.0% |
| LiveBench Mathematics | LiveBench | 96.8% | 96.0%† |
| Arena Maths | Arena (LMArena) | 1487 | 1510 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 75.6% | – |
| LiveBench 2.0 Data Analysis | LiveBench | 83.0%† | 79.6%† |
| LiveBench Research | LiveBench | 85.2%† | 68.6%† |
| LiveBench 1.0 Data Analysis | LiveBench | 83.0% | 79.6%† |
| AA-Omniscience | Artificial Analysis | 43.4 | 25 |
| PRBench Finance | Scale AI SEAL | 47.5%† | 59.5%† |
| PRBench Legal | Scale AI SEAL | 48.4%† | 61.6%† |
| BioMysteryBench | Vals AI | 79.3% | – |
| Excel Modeling Benchmark | Vals AI | 71.7% | 67.4% |
| Excel Modeling Benchmark v1.1 | Vals AI | 72.7% | – |
| MedCode | Vals AI | 48.5% | – |
| MedScribe | Vals AI | 87.9% | – |
| DrugDiscoveryBench | Scale AI SEAL | 68.7% | 62.2%† |
| EBR-bench | Epoch AI Benchmarking Hub | 76.2% | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 34.2% | 27.6%† |
| AA-Omniscience: accuracy | Artificial Analysis | 62.6% | 43.6% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 48.7% | 67.1% |
| Instruction following | |||
| LiveBench Language | LiveBench | 89.4% | 82.8%† |
| LiveBench Instruction Following | LiveBench | 75.6% | 78.0%† |
| Arena Instruction Following | Arena (LMArena) | 1473 | 1485 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1477 | 1496 |
| Arena Creative Writing | Arena (LMArena) | 1450 | 1463 |
| Arena Multi-turn | Arena (LMArena) | 1485 | 1494 |
| Arena Longer Queries | Arena (LMArena) | 1489 | 1497 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1280 | 1308 |
| LiveBench Design | LiveBench | 82.0%† | 74.8%† |
| MMMU-Pro | Artificial Analysis | 86.9% | 82.0%† |
| SAGE | Vals AI | 46.4% | – |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 49.7%† | – |
| AA-3D | Artificial Analysis | 1166 | 1037 |
| Long context | |||
| AA-LCR | Artificial Analysis | 80.7% | 83.0% |
| Arena Document | Arena (LMArena) | 1473 | 1468 |
| MLCR | Artificial Analysis | 41.7% | 13.9% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 168.8† | 157.2† |
| LiveBench 2.0 | LiveBench | 86.1%† | 74.8%† |
| LiveBench 1.0 | LiveBench | 82.2% | 81.6%† |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 52.7 | 48.1 |
| Vals Index | Vals AI | 65.0 | – |
| Vals Finance Index | Vals AI | 55.6% | – |
| Vals RSI Index | Vals AI | 27.1% | 19.6% |
| Vals Science Index | Vals AI | 65.1% | – |
| Vals Router Index | Vals AI | 65.0%† | – |
| AA Cyber Index | Artificial Analysis | 33.4 | – |
Data as of 2026-10-24. Best configuration of each model; every score links to its source on the model pages.
What changed
- GPT-6 Astra (max): 9 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
- Muse Spark 1.3 (max): 3 new results (Harvey's Legal Agent Benchmark v1, Vals RSI Index, SWE-bench Multimodal); 1 score revision.
- GPT-6 Astra (max): 1 score revision.
- Muse Spark 1.3 (max): 1 score revision.
- GPT-6 Astra (max): 4 score revisions.
- Muse Spark 1.3 (max): 1 score revision.
- GPT-6 Astra (max): BenchLeader Index 69.8 → 70.6.
- GPT-6 Astra (max): 1 score revision.
- Muse Spark 1.3 (max): 2 new results (ProgramBench, SREBench); 1 score revision.
- GPT-6 Astra (max): 1 new result (Arena Search); 3 score revisions.
Related comparisons
- GPT-6 Astra vs Claude Opus 5.5
- GPT-6 Astra vs Gemini 4 Argon
- GPT-6 Astra vs Muse Wonder Full
- Muse Spark 1.3 vs Claude Opus 5.5
- Muse Spark 1.3 vs Gemini 4 Argon
- Muse Spark 1.3 vs Muse Wonder Full
- Muse Spark 1.3 vs Claude Fable 5.1
- GPT-5.6 Sol vs Muse Spark 1.3
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
GPT-6 Astra vs Muse Spark 1.3: questions
- Is GPT-6 Astra better than Muse Spark 1.3?
- GPT-6 Astra (max) and Muse Spark 1.3 (max) are neck and neck on quality (70.4 vs 67.1, within the ±5.8 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-24, so check the category scores for your use.
- Which is cheaper, GPT-6 Astra or Muse Spark 1.3, and will that change?
- Muse Spark 1.3 is cheaper today: $2.00 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is GPT-6 Astra better than Muse Spark 1.3 for coding?
- GPT-6 Astra scores higher in coding (74 vs 65 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1775 for GPT-6 Astra against 1657 for Muse Spark 1.3.
- Is GPT-6 Astra better than Muse Spark 1.3 for agents and tool use?
- Muse Spark 1.3 scores higher in agentic tasks (66 vs 64 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 41.4% for GPT-6 Astra against 50.5% for Muse Spark 1.3.
- Which is better for long documents?
- Muse Spark 1.3 scores higher on long-context benchmarks (67 vs 65). Both take about 1.1M tokens: GPT-6 Astra 1,050,000, Muse Spark 1.3 1,048,576.
- Which has lower latency?
- Muse Spark 1.3 (max) answers first: 90.1 s against 247.0 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.
- Which is faster, GPT-6 Astra or Muse Spark 1.3?
- Muse Spark 1.3 streams faster: 116 against 58 output tokens per second.
- Do independent benchmarks agree on GPT-6 Astra vs Muse Spark 1.3?
- Across the 56 benchmarks both have, GPT-6 Astra leads 38, Muse Spark 1.3 leads 18. Independent sources split: GPT-6 Astra leads on AA Intelligence Index v4.3.2; Muse Spark 1.3 leads on Arena Text.
- When was GPT-6 Astra released?
- GPT-6 Astra was released on 2 Sept 2026.