Muse Ember vs Muse Spark Open 816B A42B
Full profiles: Muse Ember (Meta) and Muse Spark Open 816B A42B (Meta): every benchmark, price and rank.
Updated · Data from 6 sources
As of 16 Nov 2026, there are not enough shared benchmarks for a verdict, and Muse Ember is 10× cheaper at $0.200 vs $2.00 per 1M tokens blended. Muse Spark Open 816B A42B leads in every category we measure both in.
Key facts (as of 16 Nov 2026)
- Muse Ember scores 52.1 on the BenchLeader Index and Muse Spark Open 816B A42B scores 58.1 (BenchLeader, 16 Nov 2026).
- Muse Ember costs $0.200 per 1M tokens blended (list price, 16 Nov 2026).
- Muse Spark Open 816B A42B costs $2.00 per 1M tokens blended (list price, 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 | Muse Spark Open 816B A42B | 65 vs 60 category score | Medium |
| Reasoning | Muse Spark Open 816B A42B | 63 vs 51 category score | High |
| Knowledge | Muse Spark Open 816B A42B | 62 vs 49 category score | High |
| Lowest cost | Muse Ember | $0.200 vs $2.00 per 1M blended; agentic coding session $0.008 vs $0.092 | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Not enough shared benchmarks for a verdict.
- Muse Spark Open 816B A42B (xhigh) is stronger in coding, knowledge, reasoning.
- Muse Ember (max) is 10× cheaper ($0.200 vs $2.00 per 1M blended).
At a glance
| Spec | Muse Ember | Muse Spark Open 816B A42B |
|---|---|---|
| Developer | Meta | Meta |
| Released | 9 Nov 2026 | 15 Oct 2026 |
| Context window (tokens) | 1,048,576 | 1,048,576 |
| Compared at | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | $0.100 / $0.500 / $0.0040 | $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 | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 51.1 | 63.2 | Muse Spark Open 816B A42B |
| Coding | 60.0 | 65.1 | Muse Spark Open 816B A42B |
| Agents & tools | 52.4 | – | not comparable |
| Maths | – | 58.7 | not comparable |
| Knowledge | 48.8 | 61.5 | Muse Spark Open 816B A42B |
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, Long context, Multimodal, Human preference.
Where sources disagree
Across the 0 benchmarks both have, Muse Ember leads 0, Muse Spark Open 816B A42B leads 0.
| Overall index | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) | Leader |
|---|---|---|---|
| BenchLeader Index | 52.1 | 58.1 | neck and neck |
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 | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $0.200 | $2.00 |
| Input / output per 1M | $0.100 / $0.500 | $1.25 / $4.25 |
| Cached input per 1M | $0.0040 | $0.150 |
| Price per task: chat reply | $0.0002 | $0.0018 |
| Price per task: summarize a 30-page report | $0.0015 | $0.018 |
| Price per task: code edit | $0.0014 | $0.014 |
| Price per task: agentic coding session | $0.0080 | $0.092 |
| Price per task: structured extraction | $0.0003 | $0.0034 |
| Agentic coding session, input mostly cached | $0.0037 | $0.043 |
| Cost per benchmark task (AA Index) | $0.136 | – |
| Output tokens per task (AA Index) | 84k | – |
| Of which reasoning | 52k | – |
| Time per task (AA Index) | 3 min | – |
| Output speed | 437 tok/s | – |
| Time to first answer | 5.0 s | – |
| Time to a complete answer, last measured 13 Nov 2026 | 12.9 s | – |
Each model is compared at its best-ranked reasoning-effort setting. Muse Ember (max) scores 52.1 and takes 5 s to its first answer token; Muse Ember (xhigh) scores 49.1 in 4 s. Muse Spark Open 816B A42B (xhigh) scores 58.1. 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
42 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 | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 52.1 | 58.1 |
| Agents & tools score | BenchLeader | 52.4 | – |
| Coding score | BenchLeader | 60.0 | 65.1 |
| Knowledge score | BenchLeader | 48.8 | 61.5 |
| Reasoning score | BenchLeader | 51.1 | 63.2 |
| Maths score | BenchLeader | – | 58.7 |
| Blended price $/M | BenchLeader | $0.200 | $2.00 |
| Output speed | BenchLeader | 437 tok/s | – |
| Time to first answer | BenchLeader | 5.0 s | – |
| Context window | BenchLeader | 1,048,576 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 88.6% |
| SimpleBench | SimpleBench | 72.7%† | – |
| HLE-Diamond (AA) | Artificial Analysis | 25.1% | – |
| CritPt v2 hard subset (AA) | Artificial Analysis | 39.0% | – |
| Coding | |||
| SciCode | SciCode | – | 58.5% |
| Arena WebDev | Arena (LMArena) | 1609 | – |
| Vibe Code Bench v1.1 | Vals AI | 83.7% | – |
| SWE Atlas v1.1 | Scale AI SEAL | 56.0%† | – |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 61.3%† | – |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 56.7%† | – |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 50.0%† | – |
| ALE-Bench | Epoch AI Benchmarking Hub | – | 1590.1 |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 40.9% | – |
| Design Arena Website | Design Arena | 1330† | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 27.2% | – |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 30.6%† | – |
| APEX-Agents | Mercor | 64.3%† | 33.4%† |
| Arena Agent | Arena (LMArena) | 1.7 | – |
| Finance Agent v2.1 | Vals AI | 63.8% | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 25.0% | – |
| Legal Research Bench | Vals AI | 49.0% | – |
| Tax Agent Bench | Vals AI | 38.2% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 39.4% | – |
| Arena Agent Task Completion | Arena (LMArena) | -7.6 | – |
| AutomationBench | Artificial Analysis | 68.1% | – |
Show all 42 benchmarks
| Metric | Publisher | Muse Ember (max) | Muse Spark Open 816B A42B (xhigh) |
|---|---|---|---|
| Agents & tools | |||
| GDP.pdf (AA) | Artificial Analysis | 22.8% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1520 | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 20.0% | – |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | – | 69.8% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | – | 31.7% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 98.1% |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 56.6% |
| AA-Omniscience v1.1 | Artificial Analysis | 7.4 | – |
| Excel Modeling Benchmark v1.1 | Vals AI | 72.2% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 21.1% | – |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 82.6% | – |
| Long context | |||
| MLCR | Artificial Analysis | 18.3% | – |
| Composite | |||
| AA Intelligence Index v5.0.1 | Artificial Analysis | 36.0 | – |
| Vals Index | Vals AI | 60.5 | – |
| Vals Finance Index | Vals AI | 60.9% | – |
| Vals Legal Index | Vals AI | 75.7% | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 11.9%† | – |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
What changed
- Muse Ember (max): 1 new result (Arena Agent).
- Muse Spark Open 816B A42B (xhigh): BenchLeader Index 58.6 → 58.1.
- Muse Ember (max): 2 new results (MLCR, Arena WebDev).
- Muse Spark Open 816B A42B (xhigh): BenchLeader Index 57.4 → 58.6.
- Muse Spark Open 816B A42B (xhigh): 2 new results (SciCode, ALE-Bench).
- Muse Spark Open 816B A42B (xhigh): first results.
Related comparisons
- Muse Ember vs Mistral Large 4
- Muse Ember vs Gemini 3.1 Flash Lite
- Muse Ember vs Inkling Small
- Muse Spark Open 816B A42B vs Claude Fable 5.2
- Muse Spark Open 816B A42B vs Gemini 4 Carbon
- Muse Spark Open 816B A42B vs Claude Opus 5.5
- Muse Spark Open 816B A42B vs GPT-6 Astra
- Muse Ember vs Qwen3.5-27B
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Muse Ember vs Muse Spark Open 816B A42B: questions
- Is Muse Ember better than Muse Spark Open 816B A42B?
- Not enough shared benchmarks for a verdict. 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.
- Muse Ember vs Muse Spark Open 816B A42B: which is cheaper, and will that change?
- Muse Ember is cheaper today: $0.200 against $2.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Muse Ember or Muse Spark Open 816B A42B better for coding?
- Muse Spark Open 816B A42B scores higher in coding (65 vs 60 on the category index, where 50 is average).
- Which is better for long documents?
- Both accept 1,048,576 tokens of context.
- When was Muse Ember released?
- Muse Ember was released on 9 Nov 2026.