MetaMuse Spark 1.2
Best configuration ranks #52 of 610 on the BenchLeader Index at 62.6 ±6.5. Last measured 9 Aug 2026. Released 5 Aug 2026.
- Blended price
- $2.00/M
- $1.25 in · $4.25 out
- Output speed
- 219 tok/s
- First answer
- 24 s
- first token 3.92 s
- Context
- 1.0M
- Overall index63
- Reasoning70
- Coding67
- Maths54
- Knowledge77
- Long context66
- Composite79
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| xhigh | 61.3 | #68 | 219 tok/s | 24 s | $0.0018 | Agents & tools 51 · Coding 63 · Composite 60 · Human preference 70 · Knowledge 65 · Multimodal 66 · Reasoning 69 |
| defaultbest | 62.6 | #52 | 219 tok/s | 24 s | $0.0018 | Coding 67 · Composite 79 · Knowledge 77 · Long context 66 · Maths 54 · Reasoning 70 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| SimpleBench | – | 74.5%#9 | SimpleBench | |
| LMArena Hard Prompts | 1513#12 | – | LMArena | |
| LiveBench Reasoningnot in index | 90.0%#8 | – | LiveBench | |
| GPQA Diamond (AA)not in index | – | 90.4%#62 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | 45.5%#28 | Artificial Analysis |
Coding
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| SciCode | 56.4%#16 | – | SciCode | |
| WeirdML | 60.3%#47 | – | WeirdML | |
| LMArena Coding | 1533#10 | – | LMArena | |
| LMArena WebDev | 1534#31 | – | LMArena | |
| LiveBench Codingnot in index | 77.5%#26 | – | LiveBench | |
| SciCode (AA)not in index | – | 57.4%#13 | Artificial Analysis | |
| IOI | – | 49.5%#9 | Vals AI | |
| SWE-bench (Vals)not in index | 86.6%#13 | – | Vals AI |
Agents & tools
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| LMArena Agent | -1.6#28 | – | LMArena | |
| LiveBench Agentic Codingnot in index | 57.6%#16 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | 69.7%#21 | – | Vals AI |
Maths
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| ProofBench | – | 43.0%#23 | Vals AI | |
| LiveBench Mathematicsnot in index | 91.2%#19 | – | LiveBench |
Knowledge
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| SimpleQA Verified | 60.3%#12 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | 76.5%#23 | – | LiveBench | |
| AA-Omniscience | – | 27.2#26 | Artificial Analysis | |
| MMLU-Pro | 88.3%#20 | – | Vals AI | |
| LegalBench | 85.3%#20 | – | Vals AI | |
| CorpFin | 70.9%#5 | – | Vals AI | |
| TaxEval | 80.4%#1 | – | Vals AI |
Instruction following
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| LiveBench Languagenot in index | 78.6%#29 | – | LiveBench |
Human preference
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| LMArena Text | 1499#5 | – | LMArena |
Multimodal
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| LMArena Vision | 1304#12 | – | LMArena |
Long context
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| AA-LCR | – | 79.0%#71 | Artificial Analysis |
Composite
| Benchmark | xhigh | default | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 155.5#20 | Epoch AI Benchmarking Hub | |
| LiveBench | 78.0%#13 | – | LiveBench | |
| AA Intelligence Index | – | 39.8#36 | Artificial Analysis | |
| Vals Indexnot in index | 57.0#17 | – | Vals AI |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Meta | 136 tok/s | 3.92 s | $1.25 | $4.25 | 1.0M | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.150 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0018 | $0.0014 | 25.1 s |
| Summarise a 30-page report | 12,000 / 600 | $0.018 | $0.0077 | 26.5 s |
| Code edit | 6,000 / 1,500 | $0.014 | $0.0089 | 30.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.092 | $0.043 | 42.0 s |
| Structured extraction | 2,000 / 200 | $0.0034 | $0.0017 | 24.6 s |
See also
Data as of 9 Sept 2026. Compare these configurations.