Ranking 2026

Melhores LLMs Gratuitos

304 modelos de IA sem custo — open-source e planos gratuitos de API. Ordenados por AA Intelligence Index (Artificial Analysis, atualizado a cada 6h).

304
modelos gratuitos
124
open-source
180
API gratuita

Top 5 LLMs Gratuitos

#1open-source
Qwen: Qwen3.8 Max
AlibabaAlibaba
Score AA 45.4
ELO 1481
Velocidade 39 tok/s
#2open-source
Qwen: Qwen3.8 2.4T A95B
AlibabaAlibaba
Score AA 39.9
Velocidade 38 tok/s
#3open-source
DeepSeek: DeepSeek V4.1 Flash
DeepSeekDeepSeek
Score AA 39.5
Velocidade 232 tok/s
#4open-source
DeepSeek V4 Pro
DeepSeekDeepSeek
Score AA 36.0
ELO 1457
Velocidade 69 tok/s
#5
Motif 3 (Beta)
Motif Technologies
Score AA 32.3
0

Lista Completa de LLMs Gratuitos

#ModeloScore AA
1Qwen: Qwen3.8 Max45.4
2Qwen: Qwen3.8 2.4T A95B39.9
3DeepSeek: DeepSeek V4.1 Flash39.5
4DeepSeek V4 Pro36.0
5Motif 3 (Beta)32.3
6Muse Spark31.3
7Xiaomi: MiMo-V2-Pro28.6
8JT-4.1 Flash 236B A21B27.3
9Qwen: Qwen3.6 Plus27.0
10GLM-5-Turbo26.6
11Z.ai: GLM 5.126.1
12MiMo-V2-Omni-032725.1
13Xiaomi: MiMo-V2-Omni23.9
14MoonshotAI: Kimi K2.523.5
15Qwen: Qwen3.5-27B22.9
16MiniMax: MiniMax M2.522.8
17MiniMax: MiniMax M2.722.8
18MiMo-V2-Flash (Feb 2026)22.4
19G9v3-39A5B21.8
20Qwen: Qwen3 Max Thinking21.3
21MiniMax: MiniMax M2.120.9
22Qwen: Qwen3.5-35B-A3B19.3
23Gemma 4 31B19.0
24JT-35B-Flash18.7
25KAT-Coder-Pro V118.6
26Qwen: Qwen3.5 397B A17B18.4
27Qwen: Qwen3.6 35B A3B18.2
28Claude 3.7 Sonnet (thinking)17.7
29Doubao Seed Code16.9
30DeepSeek V3.216.0
31Qwen: Qwen3 Max15.6
32Qwen: Qwen3.5-122B-A10B15.6
33Gemini 2.5 Flash Preview (Sep '25) (Reasoning)15.5
34Ling 3.0 Tiny15.3
35MoonshotAI: Kimi K2 090515.3
36Gemini 2.5 Pro Preview (Mar' 25)15.0
37Z.ai: GLM 4.7 Flash14.9
38DeepSeek V3.2 Speciale14.5
39ERNIE 5.0 Thinking Preview14.3
40Grok Code Fast 114.1
41DeepSeek V3.1 Terminus13.9
42Apriel-v1.5-15B-Thinker13.8
43DeepSeek V3.113.7
44Apriel-v1.6-15B-Thinker13.4
45Command A+13.1
46DeepSeek: R113.1
47GLM-4.5 (Reasoning)12.8
48JT-MINI12.2
49Qwen: Qwen3 235B A22B Instruct 250712.0
50Sonar Reasoning Pro11.8
51Nemotron Cascade 2 30B A3B11.7
52LongCat Flash Lite11.5
53K2 Think V211.5
54HyperCLOVA X SEED Think (32B)11.4
55Grok 4.1 Fast11.3
56Z.ai: GLM 4.5 Air11.1
57Mi:dm K 2.5 Pro11.0
58Ring-1T10.9
59G9v3-3B10.8
60Arcee AI: Trinity Large Thinking10.8
61INTELLECT-310.6
62GPT-5 (ChatGPT)10.4
63Solar Open 100B (Reasoning)10.4
64Llama 4 Maverick10.0
65MiniMax M1 40k10.0
66Qwen: Qwen3 VL 235B A22B Instruct9.9
67North Mini Code9.9
68o1-mini9.8
69Qwen: Qwen3 Coder 30B A3B Instruct9.6
70Qwen: Qwen3 Next 80B A3B Instruct9.6
71GPT-4.5 (Preview)9.6
72Tri-21B-think Preview9.6
73DiffusionGemma 26B A4B9.5
74Gemini 2.0 Flash Thinking Experimental (Jan '25)9.4
75Qwen: Qwen3 Coder Next9.2
76Ling-1T9.2
77Mistral: Mistral Medium 3.19.2
78Motif-2-12.7B-Reasoning9.2
79Magistral Medium 19.1
80Solar Pro 2 (Preview) (Reasoning)9.1
81K2-V2 (medium)9.0
82Mistral: Devstral Medium9.0
83Mistral: Mistral Medium 39.0
84GPT-4o (March 2025, chatgpt-4o-latest)9.0
85Tri-21B-Think9.0
86Claude 3.5 Haiku8.9
87Gemini 2.0 Flash8.9
88Llama 3.3 Nemotron Super 49B v1 (Reasoning)8.9
89Qwen3 4B 2507 (Reasoning)8.8
90Gemini 2.0 Pro Experimental (Feb '25)8.7
91Gemini 2.5 Flash Preview (Non-reasoning)8.7
92Devstral Small (May '25)8.7
93MiniCPM5-1B (Non-reasoning)8.7
94Sonar Reasoning8.7
95Devstral 28.6
96DeepSeek V38.5
97LFM2.5-2.6B8.4
98Qwen: Qwen3 VL 32B Instruct8.4
99DeepSeek: R1 Distill Qwen 32B8.4
100Nanbeige4.1-3B8.4
101Gemini 2.0 Flash (experimental)8.2
102EXAONE 4.0 32B (Reasoning)8.2
103Magistral Small 18.2
104DeepSeek R1 0528 Qwen3 8B8.1
105Llama 4 Scout8.1
106Qwen2.5 Max8.0
107Qwen: Qwen3 VL 30B A3B Instruct7.9
108R1 Distill Llama 70B7.9
109Gemini 1.5 Pro (Sep '24)7.9
110Solar Pro 2 (Preview) (Non-reasoning)7.9
111DeepSeek R1 Distill Qwen 14B7.8
112Gemma 4 E2B (Reasoning)7.8
113Falcon-H1R-7B7.8
114Qwen2.5 72B Instruct7.7
115Llama 3.3 70B Instruct7.7
116Sonar7.7
117Step3 VL 10B7.7
118QwQ 32B-Preview7.6
119Devstral Small (Jul '25)7.6
120Baidu: ERNIE 4.5 300B A47B 7.5
121Devstral Small 27.5
122Gemini 2.0 Flash-Lite (Feb '25)7.4
123Granite 4.1 30B7.4
124NVIDIA Nemotron 3 Nano 4B7.4
125Qwen: Qwen3 VL 8B Instruct7.3
126Gemini 2.0 Flash-Lite (Preview)7.3
127Kimi Linear 48B A3B Instruct7.3
128Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)7.3
129Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)7.3
130Qwen3 4B (Reasoning)7.2
131Llama 3.1 Tulu3 405B7.2
132LFM2.5-8B-A1B7.2
133GPT-4o (ChatGPT)7.2
134Olmo 3.1 32B Think7.1
135Gemini 1.5 Flash (Sep '24)7.1
136Mistral: Pixtral Large 24117.1
137Grok 2 (Dec '24)7.1
138Qwen3 VL 4B (Reasoning)7.0
139Solar Pro 2 (Non-reasoning)7.0
140Qwen2.5 Instruct 32B6.9
141Qwen3.5 2B (Reasoning)6.9
142Llama 3.1 8B Instruct6.9
143Llama 3.1 Nemotron 70B Instruct6.9
144Grok Beta6.9
145Qwen2.5 Coder 32B Instruct6.7
146Qwen3 4B 2507 Instruct6.7
147Qwen: Qwen3 30B A3B Instruct 25076.6
148Qwen3 4B (Non-reasoning)6.6
149DeepSeek-V2.56.6
150DeepSeek-V2.5 (Dec '24)6.6
151Gemini 2.0 Flash Thinking Experimental (Dec '24)6.6
152Llama 3.1 70B Instruct6.6
153Olmo 3 32B Think6.5
154Olmo 3.1 32B Instruct6.5
155DeepSeek R1 Distill Llama 8B6.5
156Gemma 4 E2B (Non-reasoning)6.5
157Mistral: Saba6.5
158Gemini 1.5 Pro (May '24)6.4
159Llama 3.2 Instruct 90B (Vision)6.4
160R1 17766.4
161Reka Flash 36.4
162Qwen2 Instruct 72B6.3
163EXAONE 4.0 32B (Non-reasoning)6.3
164Phi-4 Mini Instruct6.3
165Grok-16.3
166Gemini 1.5 Flash-8B6.2
167AI21: Jamba Large 1.76.1
168Qwen3.5 0.8B (Reasoning)6.1
169DeepHermes 3 - Mistral 24B Preview (Non-reasoning)6.1
170OLMo 2 32B6.0
171DeepSeek-Coder-V26.0
172Gemini 1.5 Flash (May '24)5.9
173Granite 4.1 3B5.9
174LFM2-24B-A2B5.9
175Microsoft: Phi 45.9
176Qwen2.5 Coder Instruct 7B 5.8
177Gemini 1.0 Ultra5.8
178Gemma 3n E4B Instruct Preview (May '25)5.8
179Phi-3 Mini Instruct 3.8B5.8
180Phi-4 Multimodal Instruct5.8
181Mistral Large5.8
182Jamba Reasoning 3B5.7
183Qwen1.5 Chat 110B5.7
184Qwen3 VL 4B Instruct5.7
185Llama 3.2 3B Instruct5.7
186Mixtral 8x22B Instruct5.7
187MiniCPM-V 4.6 1.3B5.7
188Molmo 7B-D5.6
189OLMo 2 7B5.6
190Olmo 3 7B Think5.6
191Claude 2.15.6
192Claude 2.05.5
193DeepSeek R1 Distill Qwen 1.5B5.5
194DeepSeek-V2-Chat5.5
195Ling-mini-2.05.5
196Llama 3 70B Instruct5.5
197Qwen Chat 72B5.4
198Qwen3.5 0.8B (Non-reasoning)5.4
199PALM-25.4
200LFM 40B5.4
201Llama 3.2 11B Vision Instruct5.4
202Arctic Instruct5.4
203DBRX Instruct5.3
204DeepSeek Coder V2 Lite Instruct5.3
205DeepSeek LLM 67B Chat (V1)5.3
206Gemini 1.0 Pro5.3
207Llama 2 Chat 13B5.3
208Llama 2 Chat 70B5.3
209OpenChat 3.5 (1210)5.3
210Sarvam M (Reasoning)5.3
211Jamba 1.7 Mini5.2
212Qwen3 1.7B (Reasoning)5.2
213Granite 4.0 H 1B5.2
214LFM2 2.6B5.2
215LFM2.5-1.2B-Instruct5.2
216LFM2.5-1.2B-Thinking5.2
217Gemma 3 270M5.1
218Granite 4.0 Micro5.1
219Mistral: Mixtral 8x7B Instruct5.1
220DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)5.1
221Qwen Chat 14B5.0
222Molmo2-8B5.0
223Claude Instant5.0
224Granite 4.0 1B5.0
225Llama 65B5.0
226Qwen3 1.7B (Non-reasoning)4.9
227Gemma 3 27B4.9
228LFM2 8B A1B4.9
229Qwen3 0.6B (Non-reasoning)4.8
230Qwen3 0.6B (Reasoning)4.8
231Tiny Aya Global4.8
232Gemma 3 1B Instruct4.8
233Gemma 3 4B4.8
234Gemma 3n E2B Instruct4.8
235Granite 4.0 350M4.8
236Granite 4.0 H 350M4.8
237LFM2 1.2B4.8
238LFM2.5-VL-1.6B4.8
239Llama 3 8B Instruct4.8
240Llama 3.2 1B Instruct4.8
241Gemma 3 12B3.8
242AlfredPros: CodeLLaMa 7B Instruct Solidity—
243Qwen: Qwen2.5 7B Instruct—
244Qwen: Qwen2.5 VL 72B Instruct—
245Qwen: Qwen3 235B A22B Thinking 2507—
246Qwen: Qwen3 30B A3B Thinking 2507—
247Wan 2.1—
248Arcee AI: Trinity Mini—
249Baidu: ERNIE 4.5 21B A3B Thinking—
250Baidu: ERNIE 4.5 VL 28B A3B—
251Baidu: ERNIE 4.5 VL 424B A47B —
252ByteDance: UI-TARS 7B —
253DeepSeek: DeepSeek V4 Flash 0423—
254DeepSeek: DeepSeek V4 Flash 0731 (batch)—
255DeepSeek: DeepSeek V4 Flash Vision Exp—
256DeepSeek: DeepSeek V4 Flash Vision Exp (batch)—
257DeepSeek: DeepSeek V4 Pro 0423—
258DeepSeek: DeepSeek V4 Pro 0813 (batch)—
259DeepSeek: DeepSeek V4.1 Flash (batch)—
260EssentialAI: Rnj 1 Instruct—
261Goliath 120B—
262Gemini 3 Deep Think—
263Ideogram 4.0—
264Mi:dm K 2.5 Pro Preview—
265Magnum v4 72B—
266Llama Guard 3 8B—
267Llama Guard 4 12B—
268WizardLM-2 8x22B—
269MiniMax: MiniMax-01—
270Mistral: Codestral 2508 (batch)—
271Mistral: Ministral 3 8B 2512 (batch)—
272Mistral: Ministral 8B—
273Mistral: Mistral Large 3 2512 (batch)—
274Mistral: Mistral Medium 3.1 (batch)—
275Mistral: Mistral Medium 3.5 (batch)—
276Mistral: Mistral Nemo—
277Mistral: Mistral Small 4 (batch)—
278Mistral: Voxtral Small 24B 2507—
279MoonshotAI: Kimi K2 0711—
280MythoMax 13B—
281Nex AGI: DeepSeek V3.1 Nex N1—
282Nous: Hermes 3 405B Instruct—
283Nous: Hermes 3 70B Instruct—
284Nous: Hermes 4 405B—
285Nous: Hermes 4 70B—
286NousResearch: Hermes 2 Pro - Llama-3 8B—
287GPT-3.5 Turbo (0613)—
288GPT-4o mini Realtime (Dec '24)—
289GPT-4o Realtime (Dec '24)—
290Reka Edge—
291ReMM SLERP 13B—
292Sao10K: Llama 3 8B Lunaris—
293Sao10k: Llama 3 Euryale 70B v2.1—
294Sao10K: Llama 3.1 70B Hanami x1—
295Sao10K: Llama 3.1 Euryale 70B v2.2—
296Sao10K: Llama 3.3 Euryale 70B—
297Stable Diffusion 3.5 Large—
298Stable Video Diffusion 3D—
299Tencent: Hunyuan A13B Instruct—
300TheDrummer: Cydonia 24B V4.1—
301TheDrummer: Rocinante 12B—
302TheDrummer: Skyfall 36B V2—
303TheDrummer: UnslopNemo 12B—
304Tongyi DeepResearch 30B A3B—

Como Acessar LLMs Gratuitos

Via Interface Web (sem código)

  • ChatGPT — GPT-4o mini grátis em chat.openai.com
  • Claude.ai — Claude 3.5 Haiku grátis (limite diário)
  • Gemini — Gemini 1.5 Flash grátis em gemini.google.com
  • Mistral Le Chat — Mistral Large grátis
  • Groq — Llama 3, Mistral via API rápida grátis

Rodar Localmente (open-source)

  • Ollama — instale e execute ollama run llama3
  • LM Studio — interface gráfica para modelos locais
  • Jan.ai — app desktop com interface amigável
  • Together.ai — inferência cloud com cota gratuita generosa
  • Replicate — pay-per-use, mas com créditos iniciais

Open-Source vs. API Gratuita: Qual Escolher?

CritérioOpen-SourceAPI Gratuita
Privacidade dos dadosTotal (roda local)Compartilhados com provedor
Custo de infraestruturaGPU própria ou cloudZero (dentro do limite)
Limite de usoIlimitadoRate limit diário/mensal
Facilidade de setupMédio (requer técnica)Imediato
Customização/fine-tuningTotalLimitada ou indisponível
Qualidade máxima disponívelLlama 3.1 405BGPT-4o, Gemini Ultra

Perguntas Frequentes

Qual o melhor LLM gratuito disponível hoje?▾
Com base no AA Intelligence Index, o melhor modelo gratuito atualmente é o Qwen: Qwen3.8 Max (score 45.4), que pode ser acessado gratuitamente. Em seguida aparecem Qwen: Qwen3.8 2.4T A95B e DeepSeek: DeepSeek V4.1 Flash.
Qual a diferença entre LLM gratuito e open-source?▾
LLM gratuito significa que você pode usá-lo sem pagar — via interface web (ChatGPT free, Claude.ai, Gemini) ou via API com tier gratuito. Open-source vai além: o código e os pesos são públicos, permitindo rodar localmente, modificar e usar sem limites ou privacidade comprometida. Exemplos open-source: Llama 3, Mistral, Qwen. Exemplos gratuitos mas não open-source: GPT-4o-mini (API grátis com limite), Gemini Flash.
Como rodar um LLM open-source localmente?▾
A forma mais fácil é usar o Ollama (ollama.ai): instale, execute "ollama run llama3" e o modelo roda no seu computador. Para modelos maiores (70B+), é necessário GPU com pelo menos 40GB VRAM ou usar serviços como Groq, Together.ai e Replicate — que oferecem inferência gratuita com cota mensal.
LLMs gratuitos têm qualidade suficiente para uso profissional?▾
Sim, para muitos casos de uso. Modelos como Llama 3.1 70B, Mistral Large e Qwen 2.5 72B atingem performance comparável ou superior a GPT-3.5 e competem com GPT-4 em tarefas específicas. Para uso profissional de alto volume (produção, integração via API, SLA), a API paga geralmente oferece melhor confiabilidade e suporte.

Conteúdo relacionado:

Benchmark Open-SourceFerramentas GratuitasRanking Custo-BenefícioMelhores LLMs 2026Benchmark CompletoComo Escolher um LLM