Gemini 3.5 Flash vs Muse Spark 1.2
Full profiles: Gemini 3.5 Flash (Google) and Muse Spark 1.2 (Meta): every benchmark, price and rank.
Updated · Data from 9 sources
As of 25 Oct 2026, Gemini 3.5 Flash (high) and Muse Spark 1.2 (xhigh) are neck and neck on quality (63.1 vs 63.2 on the BenchLeader Index, inside the ±5.4 margin of error), and Muse Spark 1.2 (xhigh) costs 1.7× less, which makes it the better buy for most workloads. Pick Gemini 3.5 Flash for agents & tools (63 vs 55); pick Muse Spark 1.2 for maths (67 vs 58).
Gemini 3.5 Flash benchmarks high Google
Muse Spark 1.2 benchmarks xhigh MetaOpen in the comparison toolWhich 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 | Gemini 3.5 Flash | 60 vs 55 category score; WeirdML 62.6% vs 60.3% | Medium |
| Agents & tool use | Gemini 3.5 Flash | 63 vs 55 category score; Terminal-Bench 2.1 (Vals) 74.2% vs 69.7% | High |
| Maths | Muse Spark 1.2 | 67 vs 58 category score; Arena Maths 1491 vs 1497 | High |
| Reasoning | Muse Spark 1.2 | 69 vs 61 category score; CritPt 17.7% vs 13.1% | High |
| Knowledge | Muse Spark 1.2 | 67 vs 65 category score; AA-Omniscience 27.2 vs 21.2 | Medium |
| Long documents | Muse Spark 1.2 | 65 vs 62 category score; AA-LCR 79.0% vs 73.3% | Medium |
| Images and other media | Gemini 3.5 Flash | 67 vs 65 category score; Arena Vision 1308 vs 1304 | Medium |
| Chat quality (human preference) | Muse Spark 1.2 | 69 vs 67 on Arena votes | Medium |
| Lowest cost | Muse Spark 1.2 | $2.00 vs $3.38 per 1M blended; agentic coding session $0.092 vs $0.126 | – |
| Fastest first answer | Either | 14.2 s (Gemini 3.5 Flash (high)) vs 17.6 s (Muse Spark 1.2 (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
- Gemini 3.5 Flash (high) and Muse Spark 1.2 (xhigh) are neck and neck on quality (63.1 vs 63.2, within the ±5.4 margin).
- Gemini 3.5 Flash (high) is stronger in agents & tools, coding, instruction following, multimodal.
- Muse Spark 1.2 (xhigh) is stronger in composite, human preference, knowledge, long context, maths, reasoning.
- Muse Spark 1.2 (xhigh) is 1.7× cheaper ($2.00 vs $3.38 per 1M blended).
- Muse Spark 1.2 (xhigh) streams 1.4× faster (307 vs 222 tokens per second).
At a glance
| Spec | Gemini 3.5 Flash | Muse Spark 1.2 |
|---|---|---|
| Developer | Meta | |
| Released | 19 May 2026 | 5 Aug 2026 |
| Context window (tokens) | 1,048,576 | 1,048,576 |
| Compared at | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $1.50 / $9.00 / $0.150 | $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 | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 61.3 | 69.3 | Muse Spark 1.2 |
| Coding | 59.8 | 55.1 | Gemini 3.5 Flash |
| Agents & tools | 63.4 | 55.4 | Gemini 3.5 Flash |
| Maths | 58.3 | 67.3 | Muse Spark 1.2 |
| Knowledge | 65.2 | 66.7 | Muse Spark 1.2 |
| Instruction following | 74.3 | – | not comparable |
| Human preference | 66.8 | 68.6 | Muse Spark 1.2 |
| Multimodal | 66.6 | 65.2 | Gemini 3.5 Flash |
| Long context | 61.6 | 64.5 | Muse Spark 1.2 |
| Composite | 65.9 | 73.8 | Muse Spark 1.2 |
Where sources disagree
Across the 61 benchmarks both have, Gemini 3.5 Flash leads 22, Muse Spark 1.2 leads 37 and 2 are tied. Every overall index that rates both puts Muse Spark 1.2 ahead.
| Overall index | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) | Leader |
|---|---|---|---|
| BenchLeader Index | 63.1 | 63.2 | neck and neck |
| AA Intelligence Index v4.3.2 | 32.6 | 39.6 | Muse Spark 1.2 |
| Arena Text | 1477 | 1491 | Muse Spark 1.2 |
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 | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $3.38 | $2.00 |
| Input / output per 1M | $1.50 / $9.00 | $1.25 / $4.25 |
| Cached input per 1M | $0.150 | $0.150 |
| Price per task: chat reply | $0.0033 | $0.0018 |
| Price per task: summarize a 30-page report | $0.023 | $0.018 |
| Price per task: code edit | $0.022 | $0.014 |
| Price per task: agentic coding session | $0.126 | $0.092 |
| Price per task: structured extraction | $0.0048 | $0.0034 |
| Cost to run a full benchmark suite | $3.69 | $0.975 |
| Output speed | 222 tok/s | 307 tok/s |
| Time to first answer | 14.2 s | 17.6 s |
| Time to a complete answer | 16.4 s | 19.2 s |
Per token is not per task: on Artificial Analysis' full suite Gemini 3.5 Flash (high) cost $3.69 and Muse Spark 1.2 (xhigh) $0.975, because Gemini 3.5 Flash (high) used about 2.2× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. Gemini 3.5 Flash (high) scores 63.1 and takes 14 s to its first answer token; Gemini 3.5 Flash (medium) scores 62.8 in 12 s, Gemini 3.5 Flash (minimal) scores 56.4 in 1 s. Muse Spark 1.2 (xhigh) scores 63.2 and takes 18 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
121 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 | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 63.1 | 63.2 |
| Agents & tools score | BenchLeader | 63.4 | 55.4 |
| Coding score | BenchLeader | 59.8 | 55.1 |
| Composite score | BenchLeader | 65.9 | 73.8 |
| Human preference score | BenchLeader | 66.8 | 68.6 |
| Instruction following score | BenchLeader | 74.3 | – |
| Knowledge score | BenchLeader | 65.2 | 66.7 |
| Long context score | BenchLeader | 61.6 | 64.5 |
| Maths score | BenchLeader | 58.3 | 67.3 |
| Multimodal score | BenchLeader | 66.6 | 65.2 |
| Reasoning score | BenchLeader | 61.3 | 69.3 |
| Blended price $/M | BenchLeader | $3.38 | $2.00 |
| Output speed | BenchLeader | 222 tok/s | 307 tok/s |
| Time to first answer | BenchLeader | 14.2 s | 17.6 s |
| Context window | BenchLeader | 1,048,576 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 92.8% | – |
| SimpleBench | SimpleBench | 76.7%† | 74.5%† |
| Arena Hard Prompts | Arena (LMArena) | 1495 | 1512 |
| LiveBench 2.0 Reasoning | LiveBench | – | 90.0%† |
| LiveBench 1.0 Reasoning | LiveBench | 82.0% | 90.0% |
| GPQA Diamond (AA) | Artificial Analysis | 92.2% | 90.4% |
| Humanity's Last Exam (AA) | Artificial Analysis | 42.7% | 45.5% |
| GPQA Diamond (Vals) | Vals AI | 92.7% | – |
| EnigmaEval | Scale AI SEAL | 25.4% | – |
| ARC-AGI-1 | ARC Prize | 92.5% | – |
| ARC-AGI-2 | ARC Prize | 72.1% | – |
| CritPt | Artificial Analysis | 13.1% | 17.7% |
| Chess Puzzles | Epoch AI Benchmarking Hub | 45.0% | 27.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 32.0% | – |
| LMCA | Epoch AI Benchmarking Hub | 47.1% | 48.4% |
| DTBench | Epoch AI Benchmarking Hub | 94.7% | 94.7% |
| ForecastBench | ForecastBench | 64.3%† | – |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | 79.3% | – |
| SciCode | SciCode | 53.9% | 57.4% |
| WeirdML | WeirdML | 62.6% | 60.3% |
| Arena Coding | Arena (LMArena) | 1508 | 1541 |
| Arena WebDev | Arena (LMArena) | 1499 | 1533 |
| LiveBench 2.0 Coding | LiveBench | – | 78.2%† |
| LiveBench 1.0 Coding | LiveBench | 78.2% | 77.5% |
| LiveCodeBench | Vals AI | 87.6% | – |
Show all 121 benchmarks
| Metric | Publisher | Gemini 3.5 Flash (high) | Muse Spark 1.2 (xhigh) |
|---|---|---|---|
| Coding | |||
| IOI | Vals AI | – | 21.8% |
| SWE-bench (Vals) | Vals AI | 78.8% | 86.6% |
| Code Migration | Vals AI | 26.8% | 29.9% |
| ProgramBench | Vals AI | 0.0% | 0.5% |
| Vibe Code Bench v1.1 | Vals AI | 48.7% | 79.1% |
| Surface Evolver Bench | Epoch AI Benchmarking Hub | 58.1%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 36.1% | 54.9% |
| ALE-Bench | Epoch AI Benchmarking Hub | 911.0 | – |
| FrontierSWE | Epoch AI Benchmarking Hub | – | 12.0% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 6.6% | 7.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 78.7% | 80.2% |
| Startrise Frontend | Startrise | – | 70.3%† |
| Agents & tools | |||
| APEX-Agents | Mercor | 27.5%† | 45.4%† |
| Arena Agent | Arena (LMArena) | – | -7 |
| LiveBench Long Agentic | LiveBench | – | 21.3%† |
| LiveBench Browser | LiveBench | – | 36.0%† |
| LiveBench Agentic Coding | LiveBench | 49.0% | 57.6% |
| Terminal-Bench Hard | Artificial Analysis | 40.9% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 95.3% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 74.2% | 69.7% |
| MCP Atlas | Scale AI SEAL | 83.6% | – |
| HiL-Bench | Scale AI SEAL | 27.7%† | – |
| τ³-Banking (AA) | Artificial Analysis | 32.2% | 34.9% |
| ITBench SRE (AA) | Artificial Analysis | 40.4% | – |
| Analyst Agent (AA) | Artificial Analysis | 45.0% | – |
| APEX-Agents (AA) | Artificial Analysis | 47.0% | – |
| CyberBench | Vals AI | – | 69.5% |
| Finance Agent v2 | Vals AI | 57.9% | 60.6% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 2.5% | 25.4% |
| Legal Research Bench | Vals AI | 30.8% | 43.8% |
| SkillsBench | Vals AI | 52.7% | 53.0% |
| SREBench | Vals AI | – | 0.0% |
| Tax Agent Bench | Vals AI | 21.6% | 32.8% |
| Terminal-Bench 2.0 (Vals) | Vals AI | 67.4% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 6.1% | 6.1% |
| Terminal-Bench Science | Vals AI | 5.7% | – |
| Web Search Index | Vals AI | 37.0% | – |
| GBAEval | Epoch AI Benchmarking Hub | 6.7%† | – |
| Vending-Bench 2 | Andon Labs | 5396.4† | – |
| AutomationBench | Artificial Analysis | 42.1% | 40.6% |
| GDP.pdf (AA) | Artificial Analysis | 19.8% | 17.4% |
| EnterpriseOps-Gym | Artificial Analysis | 50.1% | 47.3% |
| AA-Briefcase v1.1 | Artificial Analysis | 874 | 1327 |
| FACTS Search | 64.3%† | – | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 62.8% | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 26.8% | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 95.6% | – |
| ProofBench | Vals AI | 31.0% | 43.0%† |
| LiveBench Mathematics | LiveBench | 88.2% | 91.2% |
| AIME 2026 | MathArena | 95.0%† | – |
| HMMT February 2026 | MathArena | 95.5%† | – |
| MathArena Apex | MathArena | 32.3%† | – |
| Arena Maths | Arena (LMArena) | 1497 | 1491 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 66.2% | 60.3% |
| LiveBench 2.0 Data Analysis | LiveBench | – | 76.5%† |
| LiveBench Research | LiveBench | – | 67.7%† |
| LiveBench 1.0 Data Analysis | LiveBench | 64.9% | 76.5% |
| AA-Omniscience | Artificial Analysis | 21.2 | 27.2 |
| MMLU-Pro | Vals AI | 89.5% | 88.3% |
| LegalBench | Vals AI | 83.6% | 85.3% |
| CorpFin | Vals AI | 64.7% | 70.9% |
| TaxEval | Vals AI | 74.4% | 80.4% |
| BioMysteryBench | Vals AI | – | 64.8% |
| Excel Modeling Benchmark | Vals AI | 63.5% | 57.0% |
| MedCode | Vals AI | 55.8% | 49.4% |
| MedScribe | Vals AI | 76.6% | 90.1% |
| Public Benefits Bench v1.1 | Vals AI | 59.5% | 68.5% |
| DrugDiscoveryBench | Scale AI SEAL | 50.0% | – |
| EBR-bench | Epoch AI Benchmarking Hub | 4.8% | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 14.0%† | 16.0% |
| AA-Omniscience: accuracy | Artificial Analysis | 51.4% | 45.4% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 37.8% | 66.7% |
| Instruction following | |||
| LiveBench Language | LiveBench | 84.6% | 78.6% |
| LiveBench Instruction Following | LiveBench | 75.6% | 74.3% |
| IFBench | Artificial Analysis | 76.3% | – |
| Arena Instruction Following | Arena (LMArena) | 1465 | 1472 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1477 | 1491 |
| EQ-Bench 4 | EQ-Bench | 1081† | – |
| Arena Creative Writing | Arena (LMArena) | 1464 | 1439 |
| Arena Multi-turn | Arena (LMArena) | 1480 | 1511 |
| Arena Longer Queries | Arena (LMArena) | 1482 | 1488 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1308 | 1304 |
| LiveBench Design | LiveBench | – | 67.1%† |
| MMMU-Pro | Artificial Analysis | 84.3% | – |
| MMMU-Pro (Vals) | Vals AI | 88.3% | 86.1% |
| MortgageTax | Vals AI | 68.1% | 65.4% |
| SAGE | Vals AI | 49.9% | 47.7% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 33.6%† | – |
| Long context | |||
| AA-LCR | Artificial Analysis | 73.3% | 79.0% |
| Arena Document | Arena (LMArena) | 1453 | – |
| MLCR | Artificial Analysis | 10.0% | 7.2% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 154.9† | 155.1† |
| LiveBench 2.0 | LiveBench | – | 62.4%† |
| LiveBench 1.0 | LiveBench | 74.6% | 78.0% |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 32.6 | 39.6 |
| Vals Multimodal Index | Vals AI | 62.9% | 69.8%† |
Data as of 2026-10-25. Best configuration of each model; every score links to its source on the model pages.
What changed
- Muse Spark 1.2 (xhigh): 1 new result (AA-Briefcase v1.1).
- Gemini 3.5 Flash (high): 1 new result (Harvey's Legal Agent Benchmark v1).
- Muse Spark 1.2 (xhigh): 1 new result (Harvey's Legal Agent Benchmark v1).
- Gemini 3.5 Flash (high): 1 score revision.
- Muse Spark 1.2 (xhigh): 2 score revisions.
- Gemini 3.5 Flash (high): BenchLeader Index 62.5 → 63.2.
- Muse Spark 1.2 (xhigh): BenchLeader Index 62.6 → 63.3.
- Gemini 3.5 Flash (high): 3 new results (GDPval-AA v2.1, AA-Omniscience: accuracy, AA-Omniscience: non-hallucination).
- Muse Spark 1.2 (xhigh): 4 new results (AutomationBench, GDP.pdf (AA), MLCR…).
- Gemini 3.5 Flash (high): 1 score revision.
Related comparisons
- Gemini 3.5 Flash vs Claude Opus 5.5
- Gemini 3.5 Flash vs Gemini 4 Argon
- Gemini 3.5 Flash vs Muse Wonder Full
- Muse Spark 1.2 vs Claude Opus 5.5
- Muse Spark 1.2 vs Gemini 4 Argon
- Muse Spark 1.2 vs Muse Wonder Full
- Muse Spark 1.2 vs Claude Fable 5.1
- Gemini 3.5 Flash vs Muse Spark 1.1
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Gemini 3.5 Flash vs Muse Spark 1.2: questions
- Is Gemini 3.5 Flash better than Muse Spark 1.2?
- Gemini 3.5 Flash (high) and Muse Spark 1.2 (xhigh) are neck and neck on quality (63.1 vs 63.2, within the ±5.4 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-25, so check the category scores for your use.
- Which is cheaper, Gemini 3.5 Flash or Muse Spark 1.2, and will that change?
- Muse Spark 1.2 is cheaper today: $2.00 against $3.38 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Gemini 3.5 Flash better than Muse Spark 1.2 for coding?
- Gemini 3.5 Flash scores higher in coding (60 vs 55 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1499 for Gemini 3.5 Flash against 1533 for Muse Spark 1.2.
- Is Gemini 3.5 Flash better than Muse Spark 1.2 for agents and tool use?
- Gemini 3.5 Flash scores higher in agentic tasks (63 vs 55 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 32.2% for Gemini 3.5 Flash against 34.9% for Muse Spark 1.2.
- Which is better for long documents?
- Muse Spark 1.2 scores higher on long-context benchmarks (65 vs 62). Both accept 1,048,576 tokens of context.
- Which has lower latency?
- Gemini 3.5 Flash (high) answers first: 14.2 s against 17.6 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, Gemini 3.5 Flash or Muse Spark 1.2?
- Muse Spark 1.2 streams faster: 307 against 222 output tokens per second.
- Do independent benchmarks agree on Gemini 3.5 Flash vs Muse Spark 1.2?
- Across the 61 benchmarks both have, Gemini 3.5 Flash leads 22, Muse Spark 1.2 leads 37 and 2 are tied. Every overall index that rates both puts Muse Spark 1.2 ahead.
- When was Muse Spark 1.2 released?
- Muse Spark 1.2 was released on 5 Aug 2026.