Toolathlon-Verified
A verified tool-use benchmark variant for completing multi-step workflows with external tools.
Top models on Toolathlon-Verified — November 13, 2026
As of November 13, 2026, Claude Opus 5 leads the Toolathlon-Verified leaderboard with 80.6% , followed by GLM-5.3-Flash (78.4%) and Claude Fable 5.1 (77.8%).
Claude Opus 5
Anthropic
Reported by AnthropicGLM-5.3-Flash
Z.AI
Reported by Z.AIClaude Fable 5.1
Anthropic
Reported by Anthropic23 modelsAgentic3% of Agentic reference weightCurrentUpdated November 13, 2026
Leaderboard (23 models)
All 23 results on this page were reported by the labs that built the models.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Claude Opus 5AnthropicReported by Anthropic | 80.6% | Reported by Anthropic | Not reported | Closed |
| 2 | GLM-5.3-FlashZ.AIReported by Z.AI | 78.4% | Reported by Z.AI | Not reported | Open |
| 3 | Claude Fable 5.1AnthropicReported by Anthropic | 77.8% | Reported by Anthropic | Not reported | Closed |
| 4 | Claude Sonnet 5.5AnthropicReported by Anthropic | 77.8% | Reported by Anthropic | Not reported | Closed |
| 5 | Claude Opus 5.5AnthropicReported by Anthropic | 77.8% | Reported by Anthropic | Not reported | Closed |
| 6 | MiMo-V2.6-ProXiaomiReported by Xiaomi | 76.9% | Reported by Xiaomi | Not reported | Open |
| 7 | ERNIE 5.5 Flash PreviewBaiduReported by Baidu | 75.0% | Reported by Baidu | Not reported | Pending |
| 8 | DeepSeek V4 Pro 0813DeepSeekReported by DeepSeek | 74.1% | Reported by DeepSeek | Not reported | Open |
| 9 | Step 5 PreviewStepFunReported by StepFun | 74.1% | Reported by StepFun | Not reported | Open |
| 10 | Hy4 previewTencentReported by Tencent | 74.1% | Reported by Tencent | Not reported | Open |
| 11 | MiMo-V2.6-FlashXiaomiReported by Xiaomi | 73.6% | Reported by Xiaomi | Not reported | Open |
| 12 | Qwen3.8-Flash-NextAlibabaReported by Alibaba | 73.5% | Reported by Alibaba | Not reported | Open |
| 13 | Kimi K3Moonshot AIReported by Moonshot AI | 73.2% | Reported by Moonshot AI | Not reported | Open |
| 14 | GLM-5.3Z.AIReported by Z.AI | 73.0% | Reported by Z.AI | Not reported | Open |
| 15 | Muse Spark OpenMetaReported by Meta | 72.8% | Reported by Meta | Not reported | Open |
| 16 | Qwen3.8 MaxAlibabaReported by Alibaba | 72.5% | Reported by Alibaba | Not reported | Open |
| 17 | Ornith-1.5-397BOrnith AIReported by Ornith AI | 71.2% | Reported by Ornith AI | Not reported | Open |
| 18 | DeepSeek V4 Flash 0731DeepSeekReported by DeepSeek | 70.3% | Reported by DeepSeek | Not reported | Open |
| 19 | dots3-note PreviewDots StudioReported by Dots Studio | 55.6% | Reported by Dots Studio | Not reported | Open |
| 20 | Inkling-SmallThinking Machines LabReported by Thinking Machines Lab | 54.4% | Reported by Thinking Machines Lab | Not reported | Open |
| 21 | Laguna S 2.1PoolsideReported by Poolside | 49.7% | Reported by Poolside | Not reported | Open |
| 22 | Ornith-1.5-35B-A3BOrnith AIReported by Ornith AI | 48.7% | Reported by Ornith AI | Not reported | Open |
| 23 | Ornith-1.5-9BOrnith AIReported by Ornith AI | 41.2% | Reported by Ornith AI | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
According to BenchLM.ai, Claude Opus 5 leads the Toolathlon-Verified benchmark with a score of 80.6%, followed by GLM-5.3-Flash (78.4%) and Claude Fable 5.1 (77.8%). The top models are clustered within 2.8 points, suggesting this benchmark is nearing saturation for frontier models.
23 models have been evaluated on Toolathlon-Verified. The benchmark falls in the Agentic category. BenchAlign v5.8 gives Toolathlon-Verified 3% of the Agentic reference weight, so it moves the Agentic leaderboard and the overall ranking. Reference weights are relative weights in the calibrated model, not fixed shares of a score.
About Toolathlon-Verified
Year
2026
Tasks
Verified multi-tool workflows
Format
Interactive tool-use score
Difficulty
Advanced tool use
BenchLM keeps Toolathlon-Verified separate from the broader Toolathlon key so provider launch values do not collapse distinct benchmark variants.
Freshness and provenance
Version
Toolathlon-Verified 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Toolathlon-Verified measure?
A verified tool-use benchmark variant for completing multi-step workflows with external tools.
Which model scores highest on Toolathlon-Verified?
Claude Opus 5 by Anthropic currently leads with a score of 80.6% on Toolathlon-Verified.
How many models are evaluated on Toolathlon-Verified?
23 AI models have published results on Toolathlon-Verified in the BenchLM catalog.
Compare top models on Toolathlon-Verified
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.