UnifyBench

RankingEffort benchmarks

Muse Spark 1.2 · XHigh

Compare models

60 measured results · 4 effort settings · 58 benchmark/harness combinations

Model evidence and sources →

Change model

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare with

Capability profile

One configuration, seven capabilities

Comparison score · 0–100
Muse Spark 1.2 · XHigh capability radarComparison scores from 0 to 100. Unknown capabilities have no point; lines stop at gaps. Hollow points are preliminary. Exact values and support labels follow the chart.AgenticHard reasoningCodingHuman prefKnowledgeMultimodalLong context50100Muse Spark 1.2 · XHigh · Agentic: 53.7 · SupportedMuse Spark 1.2 · XHigh · Hard reasoning: 48.0 · SupportedMuse Spark 1.2 · XHigh · Coding: 51.5 · SupportedMuse Spark 1.2 · XHigh · Knowledge: 46.9 · SupportedMuse Spark 1.2 · XHigh · Long context: 56.6 · Supported

Gaps are unknown, not zero. Hollow points are preliminary.

  • Agentic

    A53.7Supported

  • Hard reasoning

    A48.0Supported

  • Coding

    A51.5Supported

  • Human pref

    Unknown

  • Knowledge

    A46.9Supported

  • Multimodal

    Unknown

  • Long context

    A56.6Supported

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

Muse Spark 1.2 · XHigh: 45.6 overall comparison score. Meets the overall evidence requirements.

5 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?
  • Meta · 22.5% direct comparison weight · provider report
  • Artificial Analysis · 18.7% direct comparison weight · independent of the model provider
  • Datacurve · 18.3% direct comparison weight · independent of the model provider
  • Vals AI · 16.9% direct comparison weight · independent of the model provider
  • LiveBench · 16.2% direct comparison weight · independent of the model provider
  • Epoch AI · 7.4% direct comparison weight · independent of the model provider
Which benchmark families carry the weight?
  • DeepSWE · 18.3% direct comparison weight
  • AutomationBench · 9.1% direct comparison weight
  • DeepSearchQA · 7.5% direct comparison weight
  • MRCR · 7.5% direct comparison weight
  • simpleqa · 7.4% direct comparison weight
  • Other families (26) · 50.3% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

DeepSWE v1.1 review: Epoch AI reported grading errors in at least 23 of 113 tasks. These published results remain included. Review their limitations in Epoch's review when interpreting this evidence.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-20.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed sourceRank before → afterRank movementCohort remaining / lost supportScore after (change)
Artificial Analysis70 → 67Up 3203 remaining / 241 lost48.8 (+3.1 points)
ARC Prize73 → 74Down 1444 remaining / 0 lost49.2 (+3.6 points)
Datacurve73 → 74Down 1444 remaining / 0 lost41.6 (-4.0 points)
LiveCodeBench73 → 73No movement444 remaining / 0 lost45.6 (-0.0 points)
Scale73 → 73No movement444 remaining / 0 lost47.4 (+1.7 points)
Terminal-Bench73 → 71Up 2444 remaining / 0 lost45.8 (+0.2 points)
Cognition73 → 74Down 1444 remaining / 0 lost45.4 (-0.2 points)
LiveBench73 → 76Down 3443 remaining / 1 lost46.1 (+0.5 points)
SWE-rebench73 → 74Down 1444 remaining / 0 lost45.0 (-0.6 points)
Epoch AI72 → 80Down 8431 remaining / 13 lost44.8 (-0.9 points)
Vals AI73 → 74Down 1443 remaining / 1 lost44.6 (-1.1 points)
Provider reports · Anthropic73 → 73No movement444 remaining / 0 lost45.6 (-0.0 points)
Provider reports · Moonshot AI73 → 73No movement444 remaining / 0 lost45.2 (-0.5 points)
Provider reports · Meta73 → 68Up 5444 remaining / 0 lost51.5 (+5.9 points)
Provider reports · Mistral73 → 73No movement444 remaining / 0 lost45.6 (+0.0 points)
Provider reports · OpenAI73 → 73No movement444 remaining / 0 lost45.6 (+0.0 points)
All provider reports73 → 69Up 4444 remaining / 0 lost51.0 (+5.4 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family
Removed evidenceRecomputed scoreChange
deepswe41.6-4.0
arc-agi48.1+2.4
harvey43.8-1.8
vals-finance-agent44.2-1.5
terminal-bench47.0+1.4
vals-code-migration46.8+1.2
mmmu-pro46.8+1.2
simpleqa44.5-1.1
multi-swe-bench44.7-1.0
hle46.4+0.8
vals-excel-modeling46.4+0.7
livebench-language46.3+0.7
swe-rebench45.0-0.6
mmlu-pro46.2+0.5
livebench-instructions45.1-0.5
livebench-math46.1+0.5
vals-tax-agent46.1+0.5
omniscience46.1+0.5
epoch-game-puzzles45.3-0.4
vals-legal-research45.3-0.3
livebench-coding45.9+0.3
enterprise-ops45.4-0.3
swe-atlas-qna45.4-0.3
frontiermath45.9+0.2
tau-bench45.4-0.2
frontiercode45.4-0.2
ifbench45.4-0.2
swe-atlas-test-writing45.4-0.2
vibe-code45.4-0.2
gdpval45.5-0.2
arc-agi-345.5-0.2
gdp-pdf45.8+0.2
livebench-data45.8+0.1
briefcase45.5-0.1
gpqa45.7+0.1
automationbench45.7+0.1
aime45.7+0.1
critpt45.7+0.1
apex-agents45.7+0.1
terminal-bench-science45.7+0.0
itbench45.6-0.0
livebench-reasoning45.6-0.0
enigma-eval45.6-0.0
aa-lcr45.6-0.0
swe-atlas-refactoring45.6-0.0
livecodebench45.6-0.0
swe-bench-pro45.7+0.0
analyst-agent45.7+0.0
math50045.6+0.0
gmmlu45.6+0.0
What evidence is missing from the fit?

Across all collected settings for this model: 82 observations, 52 contributing, 0 matched without graph weight and 30 excluded. Counts do not establish rank eligibility.

  • 3 · Benchmark family or source protocol has not been reviewed
  • 1 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
  • 7 · Evaluator has not been reviewed for independent admission
  • 1 · Effort is not identified
  • 1 · Secondary GDPval-AA v2 leaderboard report; the original evaluator already supplies this exact configuration. Reprinted or rerated board snapshots are not additional evaluation evidence.
  • 11 · Supporting evidence outside the reviewed capability core
  • 3 · No matched opponent in this evaluation unit
  • 1 · Composite index or publisher-specific internal evaluation lacks a stable general-capability comparison unit.
  • 1 · Different native coding harnesses are mixed; exact shared agent configuration is not established.
  • 1 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / sourceLowHighXHighUnspecified
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis

Unit: elo. Source collected 2026-10-18.

Not reportedNot reported
1,328.29 Elo
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
79%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
40.61%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
17.71%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
EnterpriseOps-Gym-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Overall task success rate across all eight enterprise domains (oracle tool mode) · Benchmark developed by ServiceNow Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
47.27%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
17.4%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis

Unit: elo. Source collected 2026-10-18.

Not reportedNot reported
1,472.97 Elo
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
90.4%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
45.46%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Share of tasks judged accurate, complete, and concise · Expert + compound (hard) sets · Benchmark developed by Wisedocs · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
31.11%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis

Unit: index. Source collected 2026-10-18.

Not reportedNot reported
27.2
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader
Benchmark details

AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
45.38%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
57.41%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
34.85%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
Benchmark details

Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
80.15%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours
Benchmark details

Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
7.07%
Reported settings

Muse Spark 1.2 (Xhigh); Model source: https://artificialanalysis.ai/models/muse-spark-1-2

Not reported
DeepSWE v1.1Datacurve · mini-swe-agent
Benchmark details

Original public board. Configuration and task coverage may differ; inspect source.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
54.9%
Reported settings

muse-spark-1-2; Reported effort: xhigh; 113 unique tasks; 452 attempts; 4 runs

Not reported
VulcanBench v3 · Report 19VulcanBench · VulcanBench bare-bones API loop, deterministic hidden tests in Docker
Benchmark details

Frozen v3, 23 post-cutoff repository tasks. Published pass@1; repeat counts and time budgets vary by report. Display only pending comparability review.

Unit: percent. Source collected 2026-10-18.

87%
Reported settings

Muse Spark 1.2; source label: low; low effort; 23 tasks; evaluation 2026-08-25. See source for repeat counts and wall-clock budget.

73.9%
Reported settings

Muse Spark 1.2; source label: high; high effort; 23 tasks; evaluation 2026-08-25. See source for repeat counts and wall-clock budget.

52.2%
Reported settings

Muse Spark 1.2; source label: xhigh; xhigh effort; 23 tasks; evaluation 2026-08-25. See source for repeat counts and wall-clock budget.

Not reported
LiveBench · AMPS Hard · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
89%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-AMPS_Hard; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · code completion · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
80.44%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-code_completion; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · code generation · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
74.65%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-code_generation; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · connections · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
100%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-connections; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · consecutive events · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
82.55%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-consecutive_events; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · integrals with game · 2026-06-25LiveBench · LiveBench official task evaluation
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
91%
Reported settings

LiveBench official task evaluation; livebench-2026-06-25-integrals_with_game; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
LiveBench · javascript · 2026-06-25LiveBench · LiveBench Mini-SWE-Agent, 250-step limit
Benchmark details

LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only.

Unit: percent. Source collected 2026-10-18.

Not reportedNot reported
72.73%
Reported settings

LiveBench Mini-SWE-Agent, 250-step limit; livebench-2026-06-25-javascript; muse-spark-1.2-xhigh; checkpoint=muse-spark-1.2

Not reported
What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,915 results across 259 models · Collected 2026-10-18