Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

Start free brief
BenchLM

Berkeley Function Calling Leaderboard v4 (BFCL v4)

A function-calling benchmark for tool selection, schema adherence, and argument correctness.

Data verified 25 confirmed releases in the last 30 daysStart free brief

Benchmark score on BFCL v4 — August 21, 2026

We mirror the published score view for BFCL v4. Qwen3.7 Max leads the public snapshot at 75.0%, followed by BTL-4 (73.5%) and Ling 3.0 Flash (73.0%). We do not use these results to rank models overall.

1Closed

Qwen3.7 Max

Alibaba

75.0%
Overall 71.36Context 1M
2Open

BTL-4

Bad Theory Labs

73.5%
Context 262K
3Open

Ling 3.0 Flash

InclusionAI

73.0%
Overall 53.59Context 262K
13 modelsAgenticCurrentDisplay onlyUpdated August 21, 2026

Benchmark score table (13 models)

Score
1
Qwen3.7 MaxAlibaba · Closed
75.0%
2
BTL-4Bad Theory Labs · Open weight
73.5%
3
Ling 3.0 FlashInclusionAI · Open weight
73.0%
4
Qwen3.7 PlusAlibaba · Closed
72.9%
5
Pokee-Isaac 28BPokee AI · Closed
70.9%
6
LFM2.5-2.6BLiquidAI · Open weight
56.9%
7
LFM2.5-8B-A1BLiquidAI · Open weight
49.7%
8
Mellum2-12B-A2.5B-ThinkingJetBrains · Open weight
45.6%
9
Mellum2-12B-A2.5B-InstructJetBrains · Open weight
44.2%
10
ZAYA1-8BZyphra · Open weight
39.2%
11
MiniCPM5-1BOpenBMB · Open weight
25.1%
12
LFM2.5-VL-450MLiquidAI · Open weight
21.1%
13
LFM2.5-230MLiquidAI · Open weight
21.0%

The published BFCL v4 snapshot places Qwen3.7 Max first at 75.0%. The third row is 2.0 points behind. The broader top-10 range is 35.8 points, so the table still separates the published systems.

13 models have been evaluated on BFCL v4. The benchmark falls in the Agentic category. This category carries a 22% weight in BenchLM.ai's overall scoring system. BFCL v4 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About BFCL v4

Year

2026

Tasks

Function-calling tasks

Format

Tool invocation and schema evaluation

Difficulty

Advanced tool use

BenchLM stores BFCL v4 as a display-only function-calling reference outside the current weighted core schema.

Trinity-Large-Thinking: Scaling an Open Source Frontier Agent

BenchLM freshness & provenance

Version

BFCL v4 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

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.

FAQ

What does BFCL v4 measure?

A function-calling benchmark for tool selection, schema adherence, and argument correctness.

Which model scores highest on BFCL v4?

Qwen3.7 Max by Alibaba currently leads with a score of 75.0% on BFCL v4.

How many models are evaluated on BFCL v4?

13 AI models have been evaluated on BFCL v4 on BenchLM.

Compare Top Models on BFCL v4

Qwen3.7 Max vs BTL-4BTL-4 vs Ling 3.0 FlashLing 3.0 Flash vs Qwen3.7 PlusQwen3.7 Plus vs Pokee-Isaac 28B
Last updated: August 21, 2026 · BenchLM version BFCL v4 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.