Agents · Benchmark profile
A function-calling benchmark for tool selection, schema adherence, and argument correctness.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on Berkeley Function Calling Leaderboard v4
Qwen3.7-Max leads at 75%, followed by Qwen3.7-Plus (72.9%) and LFM2.5-8B-A1B (49.7%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
9 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| Qwen3.7-Max | #1 | Alibaba Cloud | 75% |
| Qwen3.7-Plus | #2 | Alibaba Cloud | 72.9% |
| LFM2.5-8B-A1B | #3 | LiquidAI | 49.7% |
| Mellum2-12B-A2.5B-Thinking | #4 | JetBrains | 45.6% |
| Mellum2-12B-A2.5B-Instruct | #5 | JetBrains | 44.2% |
| ZAYA1-8B | #6 | Zyphra | 39.2% |
| MiniCPM5-1B | #7 | OpenBMB | 25.1% |
| LFM2.5-VL-450M | #8 | LiquidAI | 21.1% |
| LFM2.5-230M | #9 | LiquidAI | 21.0% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Qwen3.7-Max qwen3.7-max | Alibaba Cloud | — | Reference only | 75% |
| #2 | Qwen3.7-Plus qwen3.7-plus | Alibaba Cloud | — | Reference only | 72.9% |
| #3 | LFM2.5-8B-A1B lfm2-5-8b-a1b | LiquidAI | open | Estimated reference | 49.7% |
| #4 | Mellum2-12B-A2.5B-Thinking mellum2-12b-a2-5b-thinking | JetBrains | open | Estimated reference | 45.6% |
| #5 | Mellum2-12B-A2.5B-Instruct mellum2-12b-a2-5b-instruct | JetBrains | open | Estimated reference | 44.2% |
| #6 | ZAYA1-8B zaya1-8b | Zyphra | open | Estimated reference | 39.2% |
| #7 | MiniCPM5-1B minicpm5-1b | OpenBMB | open | Estimated reference | 25.1% |
| #8 | LFM2.5-VL-450M lfm2-5-vl-450m | LiquidAI | open | Estimated reference | 21.1% |
| #9 | LFM2.5-230M lfm2-5-230m | LiquidAI | open | Estimated reference | 21.0% |
The top of this snapshot is led by Qwen3.7-Max at 75%; third place is 25.3 points behind. The top-9 spread is 54.0 points.
About Berkeley Function Calling Leaderboard v4
A function-calling benchmark for tool selection, schema adherence, and argument correctness. 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 function-calling benchmark for tool selection, schema adherence, and argument correctness.
Qwen3.7-Max by Alibaba Cloud currently leads with 75%.
9 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
