Tool Calling Model Ranking

See which AI models perform best on Tool Calling, which ones stay reliable, and where the biggest gaps appear. Sort by: Response Time (avg) ↓.

Models Shown

Average Tool Calling Score

8.8

Best Model

Ring-2.6-1T 10.0

Failure Reasons

With failure reason API error25 With failure reason Did not follow instructions12 With failure reason Invalid tool call11 With failure reason Wrong answer5 With failure reason No answer2

330/330

Rank	Model	Company	Tool Calling Score	Score	Total Cost	Tests Correct	Response Time (avg)
#155	Step 3.7 Flash high	Stepfun	10.0	6.8	$1.229	1/1	2.79s
Total Tests 1 Wrong Tests 0 Total Cost $1.229 Response Time (avg) 2.79s
#168	Gemini 3.5 Flash minimal	Google	10.0	6.7	$0.300	1/1	2.79s
Total Tests 1 Wrong Tests 0 Total Cost $0.300 Response Time (avg) 2.79s
#305	Elephant Alpha none	Openrouter	3.0	4.3	$0.000	0/1	2.79s
Total Tests 1 Wrong Tests 1 Total Cost $0.000 Response Time (avg) 2.79s
#246	Qwen3.8 27B none	Qwen	10.0	5.5	~$0.006	1/1	2.75s
Total Tests 1 Wrong Tests 0 Total Cost ~$0.006 Response Time (avg) 2.75s
#224	GPT-5.4 none	OpenAI	10.0	5.8	$0.397	1/1	2.75s
Total Tests 1 Wrong Tests 0 Total Cost $0.397 Response Time (avg) 2.75s
#271	Mercury 2.5 low	Inception	9.8	5.1	$0.011	1/1	2.72s
Total Tests 1 Wrong Tests 0 Total Cost $0.011 Response Time (avg) 2.72s
#256	Laguna S 2.1 high	Poolside	10.0	5.4	$0.114	1/1	2.69s
Total Tests 1 Wrong Tests 0 Total Cost $0.114 Response Time (avg) 2.69s
#300	Qwen3 Coder Next medium	Qwen	10.0	4.6	$0.034	1/1	2.64s
Total Tests 1 Wrong Tests 0 Total Cost $0.034 Response Time (avg) 2.64s
#203	Inkling low	Thinkingmachines	3.0	6.1	$0.187	0/1	2.57s
Total Tests 1 Wrong Tests 1 Total Cost $0.187 Response Time (avg) 2.57s
#280	GPT-4o-mini none	OpenAI	10.0	5.0	$0.010	1/1	2.51s
Total Tests 1 Wrong Tests 0 Total Cost $0.010 Response Time (avg) 2.51s
#264	Inkling none	Thinkingmachines	3.0	5.2	$0.147	0/1	2.50s
Total Tests 1 Wrong Tests 1 Total Cost $0.147 Response Time (avg) 2.50s
#205	Qwen3.6 Flash none	Qwen	10.0	6.1	$0.062	1/1	2.49s
Total Tests 1 Wrong Tests 0 Total Cost $0.062 Response Time (avg) 2.49s
#270	Qwen3 Coder Next none	Qwen	10.0	5.1	$0.026	1/1	2.47s
Total Tests 1 Wrong Tests 0 Total Cost $0.026 Response Time (avg) 2.47s
#272	MiMo-V2.5 none	Xiaomi	10.0	5.1	$0.025	1/1	2.43s
Total Tests 1 Wrong Tests 0 Total Cost $0.025 Response Time (avg) 2.43s
#109	Mercury 2.5 Preview medium	Inception	3.0	7.5	$0.024	0/1	2.37s
Total Tests 1 Wrong Tests 1 Total Cost $0.024 Response Time (avg) 2.37s

Tool Calling Ranking

Filter models

Top Models by Tool Calling Score

Tool Calling Score vs Total Cost

Top Models by Response Time (avg)

Tool Calling Ranking

Filter models

Top Models by Tool Calling Score

Tool Calling Score vs Total CostEstimated GPU electricity cost only. It excludes the rest of the computer and hardware ownership. NVIDIA GeForce RTX 3090: 0.0572 h × 0.300 kW × €0.2896/kWh × 1.138 USD/EUR = ~$0.005659.

Top Models by Response Time (avg)

Tool Calling Score vs Total Cost