Inteligencia de modelos de IA

Capacidad · 2026-05-12

AI Models with Tool Calling Support

Compare AI models that support tool calling / function calling for agents and automation.

¿Qué es esto?

  • Tool calling (also called function calling) lets an LLM emit a structured JSON request to invoke an external function — search, code execution, database queries, anything you expose.
  • The model returns the function name and arguments as JSON; your runtime executes it and feeds the result back as a tool message.

Por qué importa

  • Without tool calling, building agents requires brittle regex parsing of free-form text.
  • Tool calling is what makes RAG, ReAct loops and multi-step assistants reliable in production.

660 modelos con esta capacidad

ModeloEditorEntrada / 1MSalida / 1MContextoProveedores
Voxtral Small 24B 2507Mistral$0.002$0.00232K3
Meta-Llama-3.1-8B-InstructMeta$0.020$0.030128K25
Llama 3.1 8B TurboMeta$0.020$0.030131K2
Mistral Nemo Instruct 2407Mistral$0.020$0.040128K8
Hermes 4 14Bchutes$0.014$0.05441K1
Ministral 3Bazure-cognitive-services$0.040$0.040128K1
Ministral 3B (latest)Mistral$0.040$0.040128K1
Llama 3.1 8B InstructMeta$0.030$0.050131K1
Ministral 3Bazure$0.040$0.040128K1
Llama-3.2-11B-Vision-InstructMeta$0.049$0.049128K8
L3 8B Stheno V3.2novita-ai$0.050$0.0508K1
Llama 3.1 8B InstantMeta$0.050$0.080131K2
Llama 3 8BMeta$0.050$0.0808K1
Gemma 3 27BGoogle$0.027$0.109131K14
Qwen2.5-Coder 32B InstructAlibaba (Qwen)$0.027$0.109131K7
Model Routerazure-cognitive-services$0.140Unknown128K1
Model Routerazure$0.140Unknown128K1
IBM: Granite 4.1 8Bkilo$0.050$0.100131K1
DeepSeek R1 Distill Llama 70BMeta$0.030$0.130131K5
GPT OSS 20BOpenAI$0.030$0.140131K23
Amazon: Nova Micro 1.0kilo$0.035$0.140128K1
Nova Microamazon-bedrock$0.035$0.140128K1
Nova Microvercel$0.035$0.140128K1
Command R7BCohere$0.037$0.150128K2
Command R7B ArabicCohere$0.037$0.150128K1
Gemini 1.5 Flash-8BGoogle$0.037$0.1501M1
Arcee AI: Trinity Minikilo$0.045$0.150131K1
Trinity Miniclarifai$0.045$0.150131K1
GPT OSS 120BOpenAI$0.040$0.160131K33
Qwen3 235B A22B Instruct 2507Alibaba (Qwen)$0.100$0.100262K18
Qwen3-235B-A22B-Thinking-2507Alibaba (Qwen)$0.100$0.100262K17
Qwen3 30B A3B Instruct 2507Alibaba (Qwen)$0.100$0.100262K12
Qwen3-30B-A3BAlibaba (Qwen)$0.100$0.100128K9
Qwen3 30B A3B Thinking 2507Alibaba (Qwen)$0.100$0.100262K7
nvidia-nemotron-nano-9b-v2NVIDIA$0.040$0.160131K6
Qwen/Qwen3.5-9BAlibaba (Qwen)$0.050$0.150262K6
Qwen/Qwen3-VL-30B-A3B-ThinkingAlibaba (Qwen)$0.100$0.100262K6
Qwen/Qwen3-VL-30B-A3B-InstructAlibaba (Qwen)$0.100$0.100262K6
Qwen/Qwen3-VL-8B-InstructAlibaba (Qwen)$0.100$0.100262K5
deepseek-ai/DeepSeek-R1-Distill-Qwen-14BDeepSeek$0.100$0.100131K4
Reka Edgekilo$0.100$0.10016K1
Llama 3.1 8BMeta$0.100$0.10032K1
GPT OSS 120Bsynthetic$0.100$0.100128K1
Ministral 8B (latest)Mistral$0.100$0.100128K1
GLM-4.6V-FlashZ.AI / Zhipu$0.020$0.210128K3
Qwen Doc TurboAlibaba (Qwen)$0.087$0.144131K1
Mistral Small 3.2 24B InstructMistral$0.060$0.18096K3
Qwen TurboAlibaba (Qwen)$0.050$0.2001M6
nemotron-3-nano-30b-a3bNVIDIA$0.050$0.200131K6
GPT OSS 20Bfireworks-ai$0.050$0.200131K1
GPT OSS 20Bdatabricks$0.050$0.200131K1
GPT OSS Safeguard 20BOpenAI$0.070$0.200128K6
Qwen/Qwen2.5-VL-32B-InstructAlibaba (Qwen)$0.050$0.220131K6
GPT OSS 20Bfrogbot$0.070$0.200131K1
Llama-3.3-70B-InstructMeta$0.050$0.230128K22
Qwen 2.5 72B InstructAlibaba (Qwen)$0.062$0.23132K3
Mistral NemoMistral$0.150$0.150128K8
Llama 3.3 Nemotron Super 49B v1.5NVIDIA$0.050$0.250131K3
Llama 3.3 Nemotron Super 49B v1NVIDIA$0.150$0.150131K2
Pixtral 12BMistral$0.150$0.150128K2

Mostrando los 60 primeros de 660. Usa el directorio completo para filtrar más.

Frequently asked questions

How many AI models support llamada a herramientas?

660 canonical models in our database currently support llamada a herramientas. The list is regenerated on every data refresh, so it always reflects the latest model releases from models.dev.

What is the cheapest model with llamada a herramientas?

Voxtral Small 24B 2507 from Mistral is currently the lowest-priced option, at $0.002 per 1M input tokens and $0.002 per 1M output tokens. The full table above is sorted price-ascending.

Which model with llamada a herramientas has the largest context window?

Qwen Long (Alibaba (Qwen)) leads on context at 10M tokens. This may matter if you also need long-document understanding alongside llamada a herramientas.

Which models are available on the most providers?

Production-readiness usually correlates with how many independent providers host the same weights. The top three by provider count are: Kimi K2.5 (45), MiniMax-M2.5 (40), GLM-5 (38).

How is llamada a herramientas different from a regular LLM?

Tool calling lets the model emit a structured JSON request to invoke an external function (search, code execution, DB query) instead of replying with prose. Without it, agents must parse freeform text — fragile and slow.

How often is this list updated?

Daily. Our data pipeline pulls models.dev once a day, regenerates the canonical model list, and rebuilds these pages so newly released models appear within 24 hours.

Última actualización:

Prices in USD per 1M tokens. Unknown means the provider does not publish per-token pricing.

Data is sourced from models.dev and normalized for comparison. Prices and capabilities may change. Always verify critical production decisions with the provider's official documentation.