reasoning benchmark
Toolathlon Leaderboard
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
Updated Aug 17, 2026
Models37
Model coverage37
MetricScore
EvidenceB
Toolathlon Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelME | Score75.6% | Percentile100.0% | Participants37 | EvidenceC | Evaluated |
| Rank02 | ModelDE | Score74.1% | Percentile97.2% | Participants37 | EvidenceC | Evaluated |
| Rank03 | ModelMA | Score73.2% | Percentile94.4% | Participants37 | EvidenceC | Evaluated |
| Rank04 | ModelZA | Score73.0% | Percentile91.7% | Participants37 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score72.5% | Percentile88.9% | Participants37 | EvidenceC | Evaluated |
| Rank06 | ModelDE | Score70.3% | Percentile86.1% | Participants37 | EvidenceC | Evaluated |
| Rank07 | ModelAN | Score59.9% | Percentile83.3% | Participants37 | EvidenceC | Evaluated |
| Rank08 | ModelOP | Score58.0% | Percentile80.6% | Participants37 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score56.5% | Percentile77.8% | Participants37 | EvidenceC | Evaluated |
| Rank10 | ModelOP | Score55.6% | Percentile75.0% | Participants37 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score54.6% | Percentile72.2% | Participants37 | EvidenceC | Evaluated |
| Rank12 | ModelTM | Score54.4% | Percentile69.4% | Participants37 | EvidenceC | Evaluated |
| Rank13 | ModelAN | Score54.3% | Percentile66.7% | Participants37 | EvidenceC | Evaluated |
| Rank14 | ModelOP | Score53.4% | Percentile63.9% | Participants37 | EvidenceC | Evaluated |
| Rank15 | ModelOP | Score53.1% | Percentile61.1% | Participants37 | EvidenceC | Evaluated |
| Rank16 | ModelDE | Score51.8% | Percentile58.3% | Participants37 | EvidenceC | Evaluated |
| Rank17 | ModelBY | Score50.6% | Percentile55.6% | Participants37 | EvidenceC | Evaluated |
| Rank18 | ModelMA | Score50.0% | Percentile52.8% | Participants37 | EvidenceC | Evaluated |
| Rank19 | ModelPO | Score49.7% | Percentile50.0% | Participants37 | EvidenceC | Evaluated |
| Rank20 | ModelGO | Score49.4% | Percentile47.2% | Participants37 | EvidenceC | Evaluated |
| Rank21 | ModelBY | Score49.1% | Percentile44.4% | Participants37 | EvidenceC | Evaluated |
| Rank22 | ModelTE | Score48.5% | Percentile41.7% | Participants37 | EvidenceC | Evaluated |
| Rank23 | ModelZA | Score48.2% | Percentile38.9% | Participants37 | EvidenceC | Evaluated |
| Rank24 | ModelDE | Score47.8% | Percentile36.1% | Participants37 | EvidenceC | Evaluated |
| Rank25 | ModelOP | Score46.3% | Percentile33.3% | Participants37 | EvidenceC | Evaluated |
| Rank26 | ModelMI | Score46.3% | Percentile30.6% | Participants37 | EvidenceC | Evaluated |
| Rank27 | ModelDE | Score43.5% | Percentile27.8% | Participants37 | EvidenceC | Evaluated |
| Rank28 | ModelMI | Score43.5% | Percentile25.0% | Participants37 | EvidenceC | Evaluated |
| Rank29 | ModelOP | Score42.9% | Percentile22.2% | Participants37 | EvidenceC | Evaluated |
| Rank30 | ModelZA | Score40.7% | Percentile19.4% | Participants37 | EvidenceC | Evaluated |
| Rank31 | ModelAC | Score39.8% | Percentile16.7% | Participants37 | EvidenceC | Evaluated |
| Rank32 | ModelAC | Score38.3% | Percentile13.9% | Participants37 | EvidenceC | Evaluated |
| Rank33 | ModelOP | Score35.5% | Percentile11.1% | Participants37 | EvidenceC | Evaluated |
| Rank34 | ModelDE | Score35.2% | Percentile8.3% | Participants37 | EvidenceC | Evaluated |
| Rank35 | ModelDE | Score35.2% | Percentile5.6% | Participants37 | EvidenceC | Evaluated |
| Rank36 | ModelDE | Score35.2% | Percentile2.8% | Participants37 | EvidenceC | Evaluated |
| Rank37 | ModelAC | Score26.9% | Percentile0.0% | Participants37 | EvidenceC | Evaluated |
Toolathlon Highlights
The leading models and scores on this benchmark.
Toolathlon Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Toolathlon
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis toolathlon AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01MEMuse Spark 1.1Meta
- Score
- 75.6%
- Price
- $1.3 input / $4.3 output per 1M tokens
- Speed
- Up to 6.2 tok/s via Meta Model API
Strengths
- Ranks #1 of 37 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures Toolathlon, not total model capability
- 02DEDeepSeek-V4-Pro-0813DeepSeek
- Score
- 74.1%
- Price
- $0.43 input / $0.87 output per 1M tokens
- Speed
- Up to 33 tok/s via DeepSeek
Strengths
- Ranks #2 of 37 compared models
- 97th percentile on this benchmark
- C evidence result
Considerations
- This result measures Toolathlon, not total model capability
- 03MAKimi K3Moonshot AI
- Score
- 73.2%
- Price
- $3.0 input / $15 output per 1M tokens
- Speed
- Up to 26 tok/s via Fireworks
Strengths
- Ranks #3 of 37 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures Toolathlon, not total model capability
- 04ZAGLM-5.3Zhipu AI
- Score
- 73.0%
Strengths
- Ranks #4 of 37 compared models
- 92th percentile on this benchmark
- C evidence result
Considerations
- This result measures Toolathlon, not total model capability
- 05ACQwen3.8 MaxAlibaba Cloud / Qwen Team
- Score
- 72.5%
- Price
- $2.0 input / $6.0 output per 1M tokens
Strengths
- Ranks #5 of 37 compared models
- 89th percentile on this benchmark
- C evidence result
Considerations
- This result measures Toolathlon, not total model capability
Selection summary
Best AI Models for Toolathlon
Muse Spark 1.1 currently leads Toolathlon with 75.6%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is Toolathlon?
What Toolathlon measures and how its scores work.
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Toolathlon
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- toolathlon|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Toolathlon.
Which model scores highest on Toolathlon?
Muse Spark 1.1 is currently ranked first with 75.6%.
What does Toolathlon measure?
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
37 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.