Claude Fable 5 vs Muse Spark Open
As of 18 Oct 2026, Claude Fable 5 (max) leads on quality (68.8 vs 61.2 on the BenchLeader Index), but Muse Spark Open (xhigh) costs 10× less, so it is the better buy unless you need the extra quality.
Claude Fable 5 
Muse Spark Open Key differences
- Claude Fable 5 (max) leads on quality: 68.8 vs 61.2.
- Claude Fable 5 (max) is stronger in agents & tools, coding, composite, instruction following, knowledge, long context, maths, reasoning.
- Muse Spark Open (xhigh) is stronger in multimodal.
- Muse Spark Open (xhigh) is 10× cheaper ($2.00 vs $20.00 per 1M blended).
- Muse Spark Open (xhigh) streams 2.0× faster (140 vs 71 tokens per second).
Choose Claude Fable 5 (max) if
- knowledge is the priority (81 vs 73 on the category score)
- reasoning is the priority (74 vs 66 on the category score)
- long context is the priority (66 vs 64 on the category score)
Choose Muse Spark Open (xhigh) if
- cost matters: $2.00 against $20.00 per million tokens blended
- you need fast output: 140 against 71 tokens per second
- the wait for the first word matters: 15.0 s against 102.6 s to first answer
- you want to run it yourself: its weights are open
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 | Claude Fable 5 (max) | Muse Spark Open (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 73.8 | 65.5 | Claude Fable 5 |
| Coding | 73.8 | – | not comparable |
| Agents & tools | 64.9 | – | not comparable |
| Maths | 72.5 | – | not comparable |
| Knowledge | 80.9 | 72.5 | Claude Fable 5 |
| Instruction following | 63.0 | – | not comparable |
| Multimodal | – | 64.1 | not comparable |
| Long context | 66.3 | 64.2 | Claude Fable 5 |
| Composite | 79.3 | 73.1 | Claude Fable 5 |
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 | Claude Fable 5 (max) | Muse Spark Open (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $20.00 | $2.00 |
| Input / output per 1M | $10.00 / $50.00 | $1.25 / $4.25 |
| Price per task: chat reply | $0.019 | $0.0018 |
| Price per task: summarise 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 | $8.75 | $0.844 |
| Output speed | 71 tok/s | 140 tok/s |
| Time to first answer | 102.6 s | 15.0 s |
| Time to a complete answer | 109.6 s | 18.6 s |
Every benchmark
| Metric | Claude Fable 5 (max) | Muse Spark Open (xhigh) |
|---|---|---|
| BenchLeader Index | 68.8 | 61.2 |
| Agents & tools score | 64.9 | – |
| Coding score | 73.8 | – |
| Composite score | 79.3 | 73.1 |
| Instruction following score | 63.0 | – |
| Knowledge score | 80.9 | 72.5 |
| Long context score | 66.3 | 64.2 |
| Maths score | 72.5 | – |
| Reasoning score | 73.8 | 65.5 |
| Multimodal score | – | 64.1 |
| Blended price $/M | $20.00 | $2.00 |
| Output speed | 71 tok/s | 140 tok/s |
| Time to first answer | 102.6 s | 15.0 s |
| Context window | 1M | 1.0M |
| GPQA Diamond | 85.9% | – |
| FrontierMath Tiers 1–3 | 87.0% | – |
| FrontierMath Tier 4 | 90.2% | – |
| OTIS Mock AIME | 99.7% | – |
| Terminal-Bench | 44.5% | – |
| SimpleBench | 81.9% | – |
| SciCode | 61.0% | – |
| WeirdML | 91.9% | – |
| ProofBench | 95.0% | – |
| LiveBench | 83.0% | – |
| LiveBench Reasoning | 89.7% | – |
| LiveBench Coding | 86.0% | – |
| LiveBench Agentic Coding | 62.2% | – |
| LiveBench Mathematics | 96.0% | – |
| LiveBench Data Analysis | 80.5% | – |
| LiveBench Language | 90.7% | – |
| LiveBench Instruction Following | 75.8% | – |
| AA Intelligence Index v4.3.2 | 49.6 | 38.9 |
| IFBench | 63.5% | – |
| AA-LCR | 82.3% | 78.3% |
| MMMU-Pro | – | 80.9% |
| AA-Omniscience | 43.3 | 24.6 |
| Terminal-Bench Hard | 62.9% | – |
| GPQA Diamond (AA) | 92.6% | 89.0% |
| Humanity's Last Exam (AA) | 55.5% | 43.9% |
| SciCode (AA) | 61.0% | 58.5% |
| τ²-Bench Telecom (AA) | 98.5% | – |
| Vals Index | – | 46.6 |
| ARC-AGI-1 | 98.5% | – |
| ARC-AGI-2 | 89.2% | – |
| τ³-bench Banking | 39.7% | – |
| CritPt | 28.6% | 14.9% |
| GDPval-AA v2.1 | 55.5% | 45.1% |
| τ³-Banking (AA) | 38.1% | – |
| Analyst Agent (AA) | 48.8% | – |
| Code Migration | – | 21.2% |
| Excel Modeling Benchmark | – | 53.7% |
| Finance Agent v2 | – | 61.3% |
| Harvey's Legal Agent Benchmark | – | 16.7% |
| Legal Research Bench | – | 38.5% |
| Tax Agent Bench | – | 26.3% |
| Terminal-Bench 4.0 (Vals) | – | 6.6% |
| Vibe Code Bench v1.1 | – | 76.2% |
| Chess Puzzles | 47.0% | – |
| EBR-bench | 39.5% | – |
| Mystery Game Puzzles | 52.0% | – |
| PostTrainBench | 41.8% | – |
| DeepSWE v1.1 | 69.7% | – |
| LMCA | 60.3% | – |
| DTBench | 98.4% | – |
| Vending-Bench 2 | 4966.6 | – |
| GDP.pdf | 29.8% | – |
| FrontierSWE | 47.0% | – |
| Terminal-Bench 4.0 (AA) | 42.4% | 8.1% |
| Terminal-Bench 2.1 (AA) | 84.6% | 76.4% |
| AutomationBench | 54.1% | 49.7% |
| GDP.pdf | 24.0% | 18.6% |
| MLCR | 64.4% | – |
| EnterpriseOps-Gym | 51.1% | – |
| AA-Omniscience: accuracy | 65.3% | 39.3% |
| AA-Omniscience: non-hallucination | 36.4% | 75.9% |
| AA-Briefcase v1.1 | 1539 | 1279 |
| AA Openness Index | – | 27.8 |
Data as of 2026-10-18. Best configuration of each model; every score links to its source on the model pages.
Claude Fable 5 vs Muse Spark Open: questions
- Is Claude Fable 5 better than Muse Spark Open?
- Claude Fable 5 (max) leads on quality: 68.8 vs 61.2. The BenchLeader Index combines every independent quality benchmark; Claude Fable 5 (max) is ahead overall as of 2026-10-18, but check the category scores for your use.
- Which is cheaper, Claude Fable 5 or Muse Spark Open?
- Muse Spark Open is cheaper: $2.00 against $20.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Fable 5 or Muse Spark Open?
- Muse Spark Open streams faster: 140 against 71 output tokens per second.
- Which has the larger context window?
- Muse Spark Open accepts more context: 1.0M against 1M tokens.