# AI for handling customer support

- **URL canônica:** https://meirinho.com.br/en/ia/modelos/recomendar/atendimento/
- **Idioma:** en
- **Atualizado:** 2026-11-17

---
[Recommender](https://meirinho.com.br/en/ia/modelos/recomendar/)

# Best AI model to handle customer support

Support chatbots that query systems and follow company policies. Since volume is usually high, the list only considers models with input up to US$ 1 per 1M tokens.

[Describe another task](https://meirinho.com.br/en/ia/modelos/recomendar/)

## Suggested models

1. 1 [Ling 3.1 Flash](https://meirinho.com.br/en/ia/modelos/ling-3-1-flash/) Ant Group**Why it fits:** Completes 79.3% of the simulated support cases in τ-bench Verified (airline) within company policy, at a price that fits high volume.τ-bench Verified (airline): 79.3%, per [OpenRouter Benchmarks](https://openrouter.ai/benchmarks/tau2-bench-airline), latest version we checked[? what it is](https://meirinho.com.br/en/ia/modelos/glossario/tau-bench/)
2. **Price:** US$ 0.045 input and US$ 0.135 output per 1M tokens (source: [OpenRouter](https://openrouter.ai/inclusionai/ling-3.1-flash)) OpenRouter[?](https://meirinho.com.br/en/ia/modelos/glossario/preco-por-token/)
3. **Estimated cost (see the How we calculate block):** about US$ 0.0005 (estimate) OpenRouter
4. **Weights:** Undisclosed[?](https://meirinho.com.br/en/ia/modelos/glossario/pesos-abertos/)
5. **Alternative:** [MiMo V2.6 Pro](https://meirinho.com.br/en/ia/modelos/mimo-v2-6-pro/), with open weights, 78.9% on τ-bench Verified (airline)
6. 2 [MiMo V2.6 Pro](https://meirinho.com.br/en/ia/modelos/mimo-v2-6-pro/) Xiaomi**Why it fits:** Completes 78.9% of the simulated support cases in τ-bench Verified (airline) within company policy, at a price that fits high volume.τ-bench Verified (airline): 78.9%, per [OpenRouter Benchmarks](https://openrouter.ai/benchmarks/tau2-bench-airline), latest version we checked[?](https://meirinho.com.br/en/ia/modelos/glossario/tau-bench/)
7. **Price:** US$ 0.435 input and US$ 0.87 output per 1M tokens (source: [Xiaomi, official page](https://mimo.mi.com/docs/en-US/price/pay-as-you-go)) Official pricing[?](https://meirinho.com.br/en/ia/modelos/glossario/preco-por-token/)
8. **Estimated cost (see the How we calculate block):** about US$ 0.0045 (estimate) Official pricing
9. **Weights:** Open[?](https://meirinho.com.br/en/ia/modelos/glossario/pesos-abertos/)
10. **Alternative:** [Ling 3.1 Flash](https://meirinho.com.br/en/ia/modelos/ling-3-1-flash/), cheaper: US$ 0.045 input per 1M tokens, 79.3% on τ-bench Verified (airline)
11. 3 [Gemini 3.5 Flash Lite](https://meirinho.com.br/en/ia/modelos/gemini-3-5-flash-lite/) Google**Why it fits:** Completes 78.3% of the simulated support cases in τ-bench Verified (airline) within company policy, at a price that fits high volume.τ-bench Verified (airline): 78.3%, per [OpenRouter Benchmarks](https://openrouter.ai/benchmarks/tau2-bench-airline), latest version we checked[?](https://meirinho.com.br/en/ia/modelos/glossario/tau-bench/)
12. **Price:** US$ 0.30 input and US$ 2.50 output per 1M tokens On the free tier, submitted data is used to improve Google's products (source: [Google, official page](https://ai.google.dev/gemini-api/docs/pricing#gemini-3.5-flash-lite)) Official pricing[?](https://meirinho.com.br/en/ia/modelos/glossario/preco-por-token/)
13. **Estimated cost (see the How we calculate block):** about US$ 0.0046 (estimate) Official pricing
14. **Weights:** Closed[?](https://meirinho.com.br/en/ia/modelos/glossario/pesos-abertos/)
15. **Alternative:** [MiMo V2.6 Pro](https://meirinho.com.br/en/ia/modelos/mimo-v2-6-pro/), with open weights, 78.9% on τ-bench Verified (airline)
16. 4 [MiMo-V2.6-Flash](https://meirinho.com.br/en/ia/modelos/mimo-v2-6-flash/) Xiaomi**Why it fits:** Completes 78.2% of the simulated support cases in τ-bench Verified (airline) within company policy, at a price that fits high volume.τ-bench Verified (airline): 78.2%, per [OpenRouter Benchmarks](https://openrouter.ai/benchmarks/tau2-bench-airline), latest version we checked[?](https://meirinho.com.br/en/ia/modelos/glossario/tau-bench/)
17. **Price:** US$ 0.14 input and US$ 0.28 output per 1M tokens (source: [Xiaomi, official page](https://mimo.mi.com/docs/en-US/price/pay-as-you-go)) Official pricing[?](https://meirinho.com.br/en/ia/modelos/glossario/preco-por-token/)
18. **Estimated cost (see the How we calculate block):** about US$ 0.0015 (estimate) Official pricing
19. **Weights:** Open[?](https://meirinho.com.br/en/ia/modelos/glossario/pesos-abertos/)
20. **Alternative:** [Ling 3.1 Flash](https://meirinho.com.br/en/ia/modelos/ling-3-1-flash/), cheaper: US$ 0.045 input per 1M tokens, 79.3% on τ-bench Verified (airline)

Order: highest τ-bench Verified (airline) score among customer support and chatbots models, one version per model line, with input up to US$ 1.00 per 1M tokens.

### How we calculate

Change the numbers and the cost of every model on the page is recalculated right away. Without JavaScript, the default values below apply.

The full history is resent every turn, with the person's message and the model's reply. Input = turns × (turns + 1) ÷ 2 × 2 messages × words per message ÷ 0.75. Output = turns × words per message ÷ 0.75.

**With these numbers: about 8,800 input tokens and 800 output tokens (estimate).**

Conversion used: 1 token is about 4 characters or 0.75 words, OpenAI's rough rule for English text. Other languages can differ. [Conversion source: OpenAI, Understanding and counting tokens](https://help.openai.com/en/articles/4936856-what-are-tokens-and-how-to-count-them)

Fields marked as assumption are our own guesses, without a source. Adjust them to your case.

Cost of each model = input tokens × input price + output tokens × output price, with prices per 1M tokens. Models that reason before answering can use many more output tokens. It is a reference calculation and does not change the order of the list.

### Other paid models we checked

Reference paid models from the main labs. They are left out of the list above because of a lower score, no score on this metric, or not passing the case filter. Prices come from the company's official page when it publishes them.

| Model | Score on this metric | Price and source | Estimated cost |
| --- | --- | --- | --- |
| [GPT-6 Astra](https://meirinho.com.br/en/ia/modelos/gpt-6-astra/) OpenAI | 69.7% OpenRouter Benchmarks | US$ 10.00 input and US$ 50.00 output per 1M tokens (source: [OpenAI, official page](https://developers.openai.com/api/docs/pricing#latest-models)) Official pricing | about US$ 0.13 (estimate) Official pricing |
| [GPT-6.1 Sol](https://meirinho.com.br/en/ia/modelos/gpt-6-1-sol/) OpenAI | 70.7% OpenRouter Benchmarks | US$ 2.00 input and US$ 10.00 output per 1M tokens (source: [OpenAI, official page](https://developers.openai.com/api/docs/pricing#latest-models)) Official pricing | about US$ 0.026 (estimate) Official pricing |
| [GPT-6 Luna](https://meirinho.com.br/en/ia/modelos/gpt-6-luna/) OpenAI | 68.4% OpenRouter Benchmarks | US$ 0.10 input and US$ 0.50 output per 1M tokens (source: [OpenAI, official page](https://developers.openai.com/api/docs/pricing#latest-models)) Official pricing | about US$ 0.0013 (estimate) Official pricing |
| [Claude Fable 5.2](https://meirinho.com.br/en/ia/modelos/claude-fable-5-2/) Anthropic | no score on this metric | US$ 8.00 input and US$ 40.00 output per 1M tokens (source: [Anthropic, official page](https://platform.claude.com/docs/en/about-claude/pricing#model-pricing)) Official pricing | about US$ 0.10 (estimate) Official pricing |
| [Claude Opus 5.5](https://meirinho.com.br/en/ia/modelos/claude-opus-5-5/) Anthropic | 75.3% OpenRouter Benchmarks | US$ 4.00 input and US$ 20.00 output per 1M tokens (source: [Anthropic, official page](https://platform.claude.com/docs/en/about-claude/pricing#model-pricing)) Official pricing | about US$ 0.051 (estimate) Official pricing |
| [Claude Sonnet 5.5](https://meirinho.com.br/en/ia/modelos/claude-sonnet-5-5/) Anthropic | 78.8% OpenRouter Benchmarks | US$ 2.00 input and US$ 10.00 output per 1M tokens (source: [Anthropic, official page](https://platform.claude.com/docs/en/about-claude/pricing#model-pricing)) Official pricing | about US$ 0.026 (estimate) Official pricing |
| [Claude Haiku 5.5](https://meirinho.com.br/en/ia/modelos/claude-haiku-5-5/) Anthropic | 74% OpenRouter Benchmarks | US$ 0.10 input and US$ 0.50 output per 1M tokens (source: [Anthropic, official page](https://platform.claude.com/docs/en/about-claude/pricing#model-pricing)) Official pricing | about US$ 0.0013 (estimate) Official pricing |
| [Gemini 4 Argon](https://meirinho.com.br/en/ia/modelos/gemini-4-argon/) Google | 78.7% OpenRouter Benchmarks | US$ 2.00 until November 30, 2026; then US$ 4.00 input and US$ 10.00 output per 1M tokens From December 1, 2026: US$ 4.00 / US$ 20.00From December 1, 2026? (source: [Google, official page](https://aistudio.google.com/docs/pricing#gemini-4-argon)) Official pricing | about US$ 0.026 (estimate) Official pricing |
| [Gemini 3.8 Flash](https://meirinho.com.br/en/ia/modelos/gemini-3-8-flash/) Google | 76.7% OpenRouter Benchmarks | US$ 0.75 until December 31, 2026; then US$ 1.50 input and US$ 3.75 output per 1M tokens From January 1, 2027: US$ 1.50 / US$ 7.50From January 1, 2027? (source: [Google, official page](https://ai.google.dev/gemini-api/docs/pricing#gemini-3.8-flash)) Official pricing | about US$ 0.0096 (estimate) Official pricing |
| [Grok 4.7](https://meirinho.com.br/en/ia/modelos/grok-4-7/) xAI | 75.3% OpenRouter Benchmarks | US$ 2.00 input and US$ 6.00 output per 1M tokens (source: [xAI, official page](https://docs.x.ai/developers/pricing#text-api-pricing)) Official pricing | about US$ 0.022 (estimate) Official pricing |
| [Mistral Large 4](https://meirinho.com.br/en/ia/modelos/mistral-large-4/) Mistral AI | no score on this metric | US$ 0.70 input and US$ 2.10 output per 1M tokens (source: [Mistral AI, official page](https://docs.mistral.ai/inference/pricing)) Official pricing | about US$ 0.0078 (estimate) Official pricing |
| [Mistral Medium 3.5](https://meirinho.com.br/en/ia/modelos/mistral-medium-3-5/) Mistral AI | no score on this metric | US$ 1.50 input and US$ 7.50 output per 1M tokens (source: [Mistral AI, official page](https://docs.mistral.ai/models/mistral-medium-3-5-26-04)) Official pricing | about US$ 0.019 (estimate) Official pricing |
| [Command A](https://meirinho.com.br/en/ia/modelos/command-a/) Cohere | no score on this metric | US$ 2.50 input and US$ 10.00 output per 1M tokens (source: [Cohere, official page](https://docs.cohere.com/docs/command-a)) Official pricing | about US$ 0.03 (estimate) Official pricing |
| [Command A+](https://meirinho.com.br/en/ia/modelos/command-a-plus/) Cohere | no score on this metric | US$ 0.30 input and US$ 1.50 output per 1M tokens (source: [OpenRouter](https://openrouter.ai/cohere/command-a-plus); Cohere's official page does not publish this model's price; OpenRouter is kept as reference) OpenRouter | about US$ 0.0038 (estimate) OpenRouter |
| [Nova Premier 1.0](https://meirinho.com.br/en/ia/modelos/nova-premier-v1/) Amazon | no score on this metric | US$ 2.50 input and US$ 12.50 output per 1M tokens (source: [Amazon, official page](https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AmazonBedrock/current/us-east-1/index.json)) Official pricing | about US$ 0.032 (estimate) Official pricing |
| [Nova 2 Lite](https://meirinho.com.br/en/ia/modelos/nova-2-lite-v1/) Amazon | no score on this metric | US$ 0.30 input and US$ 2.50 output per 1M tokens (source: [Amazon, official page](https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AmazonBedrock/current/us-east-1/index.json)) Official pricing | about US$ 0.0046 (estimate) Official pricing |

Prices in US$ per 1 million tokens, input / output. “From” marks the price the lab has already published on its official pricing page for a future date.

[Full table for agents](https://meirinho.com.br/en/ia/modelos/uso/agentes/)

## FAQ

By this page's rule, the list starts with Ling 3.1 Flash, MiMo V2.6 Pro, Gemini 3.5 Flash Lite, MiMo-V2.6-Flash. Order: highest τ-bench Verified (airline) score among customer support and chatbots models, one version per model line, with input up to US$ 1.00 per 1M tokens.

With the default values in the How we calculate block, the estimate ranges from US$ 0.0005 to US$ 0.0046 among the listed models with a published price. It is only an estimate and you can adjust the numbers on the page.

Yes: MiMo V2.6 Pro, MiMo-V2.6-Flash.

No. The recommendation comes from fixed rules over public data from Epoch AI, OpenRouter, Hugging Face and each company's official pricing page, in the latest version we checked. The typed text is only compared with keywords.

## Terms used on this page

- [τ-bench Verified (airline)](https://meirinho.com.br/en/ia/modelos/glossario/tau-bench/)
- [Price per token](https://meirinho.com.br/en/ia/modelos/glossario/preco-por-token/)
- [Open weights](https://meirinho.com.br/en/ia/modelos/glossario/pesos-abertos/)

## Use cases with their own page

## Official pricing pages checked

- [OpenAI](https://platform.openai.com/docs/pricing), latest version we checked.
- [Anthropic](https://docs.anthropic.com/en/docs/about-claude/pricing), latest version we checked.
- [Google](https://ai.google.dev/gemini-api/docs/pricing), latest version we checked.
- [xAI](https://docs.x.ai/docs/models), latest version we checked.
- [Mistral AI](https://mistral.ai/pricing), latest version we checked.
- [Cohere](https://cohere.com/pricing), latest version we checked.
- [Amazon](https://aws.amazon.com/bedrock/pricing/), latest version we checked.

## Sources

- [Epoch AI](https://epoch.ai/benchmarks), CC BY 4.0, latest version we checked.
- [OpenRouter Benchmarks](https://openrouter.ai/benchmarks), CC BY 4.0, latest version we checked.
- [OpenRouter](https://openrouter.ai/api/v1/models), dados públicos da API; exports de benchmarks e rankings em CC BY 4.0, latest version we checked.
- [Hugging Face Hub](https://huggingface.co/api/models), metadados públicos do Hub; licença de cada modelo no repositório, latest version we checked.

Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.

[#](#)
