Agents · Benchmark profile
A Model Context Protocol task benchmark used in Qwen's launch tables to measure practical execution over MCP-style tools and integrations.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on MCP-Tasks
Qwen3.5 397B leads at 74.2%, followed by Qwen3.6 Plus (74.1%) and Claude Opus 4.5 (71.8%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
5 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| Qwen3.5 397B | #1 | Alibaba Cloud | 74.2% |
| Qwen3.6 Plus | #2 | Alibaba Cloud | 74.1% |
| Claude Opus 4.5 | #3 | Anthropic | 71.8% |
| GLM-5 | #4 | Z.AI | 60.8% |
| Kimi K2.5 | #5 | Moonshot AI | 59.1% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Qwen3.5 397B qwen3.5-397b-a17b | Alibaba Cloud | open | Estimated reference | 74.2% |
| #2 | Qwen3.6 Plus qwen3-6-plus | Alibaba Cloud | closed | Reference only | 74.1% |
| #3 | Claude Opus 4.5 claude-opus-4-5 | Anthropic | closed | Reference only | 71.8% |
| #4 | GLM-5 glm-5 | Z.AI | open | Reference only | 60.8% |
| #5 | Kimi K2.5 kimi-k2.5 | Moonshot AI | — | Reference only | 59.1% |
The top of this snapshot is led by Qwen3.5 397B at 74.2%; third place is 2.4 points behind. The top-5 spread is 15.1 points.
About MCP-Tasks
A Model Context Protocol task benchmark used in Qwen's launch tables to measure practical execution over MCP-style tools and integrations. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A Model Context Protocol task benchmark used in Qwen's launch tables to measure practical execution over MCP-style tools and integrations.
Qwen3.5 397B by Alibaba Cloud currently leads with 74.2%.
5 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
