AI API PricingCost per Token 2026

How much does it cost to use GPT, Claude, Gemini and other LLM APIs? Compare pricing per million tokens — input and output rates for every major model.

Last updated: July 17, 2026 • 614 APIs listed

614

APIs listed

188

With free tier

$0.019

Cheapest (input/1M)

$150.00

Most expensive (input/1M)

How to read the table: prices are per million tokens (input = what you send; output = the model's response). In English, 1,000 tokens ≈ 750 words ≈ 1 A4 page. Prices verified on each company's official pricing page.

Price per Million Tokens — Paid APIs

#ModelInput USD/1M
1Mistral: Mistral Nemo$0.019
2Gemma 3n E4B Instruct$0.020
3Sarvam 30B$0.026
4Qwen3.5 4B (Non-reasoning)$0.030
5Qwen3.5 4B (Reasoning)$0.030
6Granite 3.3 8B (Non-reasoning)$0.030
7Amazon: Nova Micro 1.0$0.035
8Nova Micro$0.035
9Cohere: Command R7B (12-2024)$0.037
10Qwen: Qwen2.5 7B Instruct$0.040
11HyperNova 60B 2605$0.040
12NVIDIA Nemotron Nano 9B V2 (Reasoning)$0.040
13Sao10K: Llama 3 8B Lunaris$0.040
14Sarvam 105B (high)$0.042
15Arcee AI: Trinity Mini$0.045
16Llama 3 8B Instruct$0.045
17Granite 4.1 8B$0.050
18Llama 2 Chat 7B$0.050
19NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)$0.050
20NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)$0.050
21NVIDIA Nemotron Nano 9B V2 (Non-reasoning)$0.050
22GPT-5 Nano$0.050
23GPT-5 nano (minimal)$0.050
24gpt-oss-20b$0.050
25Amazon: Nova Lite 1.0$0.060
26Nova Lite$0.060
27Gemma 3n 4B$0.060
28Granite 4.0 H Small$0.060
29MythoMax 13B$0.060
30Baidu: ERNIE 4.5 21B A3B Thinking$0.070
31GLM-4.7-Flash (Reasoning)$0.070
32ByteDance Seed: Seed 1.6 Flash$0.075
33Gemini 2.0 Flash Lite$0.075
34Mistral: Mistral Small 3.2 24B$0.075
35Nemotron 3 Nano Omni 30B A3B Reasoning$0.075
36gpt-oss-safeguard-20b$0.075
37Qwen: Qwen3 30B A3B Thinking 2507$0.080
38Tongyi DeepResearch 30B A3B$0.090
39Qwen3.5 Omni Flash$0.100
40Olmo 3 7B Instruct$0.100
41ByteDance: UI-TARS 7B $0.100
42Gemini 2.5 Flash Lite$0.100
43Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)$0.100
44Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)$0.100
45Gemma 4 12B (Reasoning)$0.100
46Ling 2.6 Flash$0.100
47Llama 3.1 8B Instruct$0.100
48Llama 3.2 1B Instruct$0.100
49Devstral Small (Jul '25)$0.100
50Ministral 3 3B$0.100
51Mistral Small 3$0.100
52Mistral Small 3.1$0.100
53Mistral Small 3.2$0.100
54Mistral: Mistral Small Creative$0.100
55Mistral: Voxtral Small 24B 2507$0.100
56GPT-4.1 Nano$0.100
57Reka Edge$0.100
58Step 3.5 Flash$0.100
59Step 3.5 Flash$0.100
60Apertus 8B Instruct$0.100
61MiMo-V2-Flash (Reasoning)$0.100
62Z.ai: GLM 4 32B $0.100
63Mistral: Mistral 7B Instruct v0.1$0.110
64Hy3-preview (Reasoning)$0.123
65Hy3-preview (Reasoning)$0.123
66Microsoft: Phi 4$0.125
67Gemma 4 26B A4B $0.130
68Nous: Hermes 4 70B$0.130
69Hermes 4 - Llama-3.1 70B (Non-reasoning)$0.130
70Hermes 4 - Llama-3.1 70B (Reasoning)$0.130
71Nex AGI: DeepSeek V3.1 Nex N1$0.135
72Baidu: ERNIE 4.5 VL 28B A3B$0.140
73DeepSeek V4 Flash$0.140
74Gemma 4 31B$0.140
75Ling-flash-2.0$0.140
76Ring-flash-2.0$0.140
77NousResearch: Hermes 2 Pro - Llama-3 8B$0.140
78Tencent: Hunyuan A13B Instruct$0.140
79MiMo-V2.5$0.140
80Qwen: Qwen3 235B A22B Thinking 2507$0.149
81EssentialAI: Rnj 1 Instruct$0.150
82Gemini 2.0 Flash$0.150
83Llama 3.2 3B Instruct$0.150
84Ministral 3 8B$0.150
85Mistral: Ministral 3 8B 2512$0.150
86Mistral: Mistral Small 4$0.150
87GPT-4o-mini (2024-07-18)$0.150
88GPT-4o-mini Search Preview$0.150
89gpt-oss-120b$0.150
90OpenAI: GPT-4o-mini$0.150
91Solar Mini$0.150
92TheDrummer: Rocinante 12B$0.170
93GLM-4.5-Air$0.170
94Llama 4 Scout$0.175
95Qwen: Qwen3 VL 8B Instruct$0.180
96Qwen3 8B (Non-reasoning)$0.180
97Qwen3 8B (Reasoning)$0.180
98Qwen3 VL 8B (Reasoning)$0.180
99Arcee AI: Spotlight$0.180
100Llama Guard 4 12B$0.180
101Jamba 1.5 Mini$0.200
102Jamba 1.6 Mini$0.200
103Qwen: Qwen3 30B A3B Instruct 2507$0.200
104Qwen: Qwen3 VL 30B A3B Instruct$0.200
105Qwen3 30B A3B (Reasoning)$0.200
106Qwen3 30B A3B 2507 (Reasoning)$0.200
107Qwen3 30B A3B 2507 Instruct$0.200
108Qwen3 VL 30B A3B (Reasoning)$0.200
109MiniMax: MiniMax-01$0.200
110Ministral 3 14B$0.200
111Mistral Small (Sep '24)$0.200
112Mistral: Ministral 3 14B 2512$0.200
113NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)$0.200
114NVIDIA Nemotron Nano 12B v2 VL (Reasoning)$0.200
115GPT-5.4 Nano$0.200
116Reka Flash 3$0.200
117Step 3.7 Flash$0.200
118Grok 4 Fast$0.200
119Seed-OSS-36B-Instruct$0.210
120Arcee AI: Trinity Large Thinking$0.220
121Trinity Large Thinking$0.235
122Qwen: Qwen2.5 VL 72B Instruct$0.250
123Qwen3 Omni 30B A3B (Reasoning)$0.250
124Qwen3 Omni 30B A3B Instruct$0.250
125Anthropic: Claude 3 Haiku$0.250
126ByteDance Seed: Seed-2.0-Lite$0.250
127Gemini 3.1 Flash Lite$0.250
128Gemini 3.1 Flash Lite Preview$0.250
129Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)$0.250
130Inception: Mercury 2$0.250
131Mistral 7B Instruct$0.250
132NVIDIA Nemotron 3 Super 120B A12B (Reasoning)$0.250
133GPT-5 Mini$0.250
134GPT-5 mini (minimal)$0.250
135GPT-5.1-Codex-Mini$0.250
136DeepSeek V3.2 Exp$0.270
137Baidu: ERNIE 4.5 300B A47B $0.280
138DeepSeek V3.2$0.280
139DeepSeek V3.2 Exp (Non-reasoning)$0.280
140DeepSeek V3.2 Exp (Reasoning)$0.280
141Amazon: Nova 2 Lite$0.300
142Nova 2.0 Lite (high)$0.300
143Nova 2.0 Omni (low)$0.300
144Nova 2.0 Omni (medium)$0.300
145Nova 2.0 Omni (Non-reasoning)$0.300
146Gemini 2.5 Flash$0.300
147Gemini 2.5 Flash Preview (Reasoning)$0.300
148Nano Banana (Gemini 2.5 Flash Image)$0.300
149Ling-2.6-1T$0.300
150Ring-2.6-1T$0.300
151Kwaipilot: KAT-Coder-Pro V2$0.300
152MiniMax-M2$0.300
153MiniMax-M3$0.300
154MiniMax: MiniMax M2-her$0.300
155MiniMax: MiniMax M2.1$0.300
156MiniMax: MiniMax M2.5$0.300
157MiniMax: MiniMax M2.7$0.300
158Mistral: Codestral 2508$0.300
159Nous: Hermes 3 70B Instruct$0.300
160TheDrummer: Cydonia 24B V4.1$0.300
161Grok 3 Mini$0.300
162Grok 3 Mini Beta$0.300
163GLM-4.6V (Non-reasoning)$0.300
164GLM-4.6V (Reasoning)$0.300
165Llama 3.2 11B Vision Instruct$0.345
166Qwen3 14B (Non-reasoning)$0.350
167Qwen3 14B (Reasoning)$0.350
168Llama 4 Maverick$0.350
169Mistral: Mistral Small 3.1 24B$0.350
170DeepSeek V3$0.360
171Qwen3.5 Omni Plus$0.400
172Qwen3.7 Plus$0.400
173MiniMax: MiniMax M1$0.400
174Mistral: Devstral 2 2512$0.400
175Mistral: Devstral Medium$0.400
176Mistral: Mistral Medium 3$0.400
177Mistral: Mistral Medium 3.1$0.400
178Llama Nemotron Super 49B v1.5 (Non-reasoning)$0.400
179Llama Nemotron Super 49B v1.5 (Reasoning)$0.400
180GPT-4.1 Mini$0.400
181TheDrummer: UnslopNemo 12B$0.400
182Baidu: ERNIE 4.5 VL 424B A47B $0.420
183DeepSeek V4 Pro$0.435
184Xiaomi: MiMo-V2.5-Pro$0.435
185Qwen: Qwen3 Coder 30B A3B Instruct$0.450
186Mistral: Mixtral 8x7B Instruct$0.450
187ReMM SLERP 13B$0.450
188Qwen2.5 72B Instruct$0.475
189Llama Guard 3 8B$0.480
190Qwen: Qwen3 Next 80B A3B Instruct$0.500
191Qwen3 Next 80B A3B (Reasoning)$0.500
192Arcee AI: Coder Large$0.500
193Command-R (Mar '24)$0.500
194Gemini 3 Flash Preview (Non-reasoning)$0.500
195Gemini 3 Flash Preview (Reasoning)$0.500
196Google: Nano Banana 2 (Gemini 3.1 Flash Image)$0.500
197Nano Banana 2 (Gemini 3.1 Flash Image Preview)$0.500
198Magistral Small 1.2$0.500
199Nex-N2-Pro$0.500
200GPT-3.5 Turbo$0.500
201GPT-3.5 Turbo$0.500
202MiniMax M1 80k$0.550
203TheDrummer: Skyfall 36B V2$0.550
204GLM-4.6 (Reasoning)$0.550
205Llama 3.1 70B Instruct$0.560
206MoonshotAI: Kimi K2 0711$0.570
207GLM-4.6 (Non-reasoning)$0.575
208Llama 3.3 70B Instruct$0.580
209Kimi K2$0.585
210DeepSeek V3.1$0.590
211Kimi K2 Thinking$0.600
212MoonshotAI: Kimi K2 0905$0.600
213MoonshotAI: Kimi K2.5$0.600
214Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)$0.600
215GPT Audio Mini$0.600
216Writer: Palmyra X5$0.600
217GLM-4.5V (Non-reasoning)$0.600
218GLM-4.5V (Reasoning)$0.600
219GLM-4.7 (Reasoning)$0.600
220WizardLM-2 8x22B$0.620
221Gemma 2 27B$0.650
222Llama 3 70B Instruct$0.650
223Sao10K: Llama 3.3 Euryale 70B$0.650
224QwQ 32B$0.660
225Nemotron 3 Ultra 550B A55B (Reasoning)$0.675
226Qwen: Qwen3 235B A22B Instruct 2507$0.700
227Qwen: Qwen3 VL 235B A22B Instruct$0.700
228Qwen: Qwen3 VL 32B Instruct$0.700
229Qwen3 235B A22B (Reasoning)$0.700
230Qwen3 32B (Non-reasoning)$0.700
231Qwen3 32B (Reasoning)$0.700
232Qwen3 VL 235B A22B (Reasoning)$0.700
233Qwen3 VL 32B (Reasoning)$0.700
234DeepSeek: R1$0.700
235R1 Distill Llama 70B$0.700
236Hermes 3 - Llama-3.1 70B$0.700
237Arcee AI: Virtuoso Large$0.750
238Mancer: Weaver (alpha)$0.750
239GPT-5.4 Mini$0.750
240AionLabs: Aion-2.0$0.800
241AionLabs: Aion-RP 1.0 (8B)$0.800
242AlfredPros: CodeLLaMa 7B Instruct Solidity$0.800
243Amazon: Nova Pro 1.0$0.800
244Nova Pro$0.800
245Claude 3.5 Haiku$0.800
246Morph: Morph V3 Fast$0.800
247Apertus 70B Instruct$0.820
248Relace: Relace Apply 3$0.850
249Sao10K: Llama 3.1 Euryale 70B v2.2$0.850
250Arcee AI: Maestro Reasoning$0.900
251Morph: Morph V3 Large$0.900
252Kimi K2.7 Code$0.950
253MoonshotAI: Kimi K2.6$0.950
254Claude 4.5 Haiku (Reasoning)$1.00
255Claude Haiku 4.5$1.00
256Mistral Small (Feb '24)$1.00
257Nous: Hermes 3 405B Instruct$1.00
258Nous: Hermes 4 405B$1.00
259Hermes 4 - Llama-3.1 405B (Non-reasoning)$1.00
260Hermes 4 - Llama-3.1 405B (Reasoning)$1.00
261GPT-5.6 Luna (high)$1.00
262GPT-5.6 Luna (low)$1.00
263GPT-5.6 Luna (max)$1.00
264GPT-5.6 Luna (medium)$1.00
265GPT-5.6 Luna (Non-reasoning)$1.00
266GPT-5.6 Luna (xhigh)$1.00
267OpenAI: GPT-5.6 Luna Pro$1.00
268Relace: Relace Search$1.00
269Grok Build 0.1 0616$1.00
270xAI: Grok Build 0.1$1.00
271GLM-5 (Non-reasoning)$1.00
272GLM-5 (Reasoning)$1.00
273o3 Mini$1.10
274o3 Mini High$1.10
275o4 Mini$1.10
276o4 Mini High$1.10
277DeepSeek V3 0324$1.14
278Qwen3 Max (Preview)$1.20
279Qwen3 Max Thinking (Preview)$1.20
280Llama 3.1 Nemotron 70B Instruct$1.20
281Nova 2.0 Pro Preview (medium)$1.25
282Cogito v2.1 (Reasoning)$1.25
283Deep Cogito: Cogito v2.1 671B$1.25
284Gemini 2.5 Pro$1.25
285Gemini 2.5 Pro Preview (May' 25)$1.25
286Gemini 2.5 Pro Preview 05-06$1.25
287Gemini 2.5 Pro Preview 06-05$1.25
288Muse Spark 1.1 (xhigh)$1.25
289GPT-5$1.25
290GPT-5 (ChatGPT)$1.25
291GPT-5 (minimal)$1.25
292GPT-5 Chat$1.25
293GPT-5 Codex$1.25
294GPT-5.1$1.25
295GPT-5.1 Chat$1.25
296GPT-5.1-Codex$1.25
297GPT-5.1-Codex-Max$1.25
298Grok 4.20 Multi-Agent$1.25
299Grok 4.3$1.25
300Qwen3.6 Max Preview$1.30
301DeepSeek R1 (Jan '25)$1.35
302Llama 3.2 Instruct 90B (Vision)$1.38
303GLM-5.1 (Non-reasoning)$1.40
304GLM-5.1 (Reasoning)$1.40
305GLM-5.2 (max)$1.40
306Sao10k: Llama 3 Euryale 70B v2.1$1.48
307Qwen3 Coder 480B A35B Instruct$1.50
308Gemini 3.5 Flash (minimal)$1.50
309Google: Gemini 3.5 Flash$1.50
310Mistral: Mistral Medium 3.5$1.50
311DeepSeek V3.1 Terminus$1.64
312GPT-5.2$1.75
313GPT-5.2-Codex$1.75
314GPT-5.3 Chat$1.75
315GPT-5.3-Codex$1.75
316Inkling$1.87
317AI21: Jamba Large 1.7$2.00
318Jamba 1.5 Large$2.00
319Jamba 1.6 Large$2.00
320Claude Sonnet 5$2.00
321Gemini 3 Pro Preview (high)$2.00
322Gemini 3 Pro Preview (low)$2.00
323Gemini 3.1 Pro Preview$2.00
324Gemini 3.1 Pro Preview Custom Tools$2.00
325Google: Nano Banana Pro (Gemini 3 Pro Image)$2.00
326Nano Banana Pro (Gemini 3 Pro Image Preview)$2.00
327Mistral Large 2 (Jul '24)$2.00
328Mistral Large 2 (Nov '24)$2.00
329Magistral Medium 1.2$2.00
330Mistral Large$2.00
331Mistral: Mixtral 8x22B Instruct$2.00
332Mistral: Pixtral Large 2411$2.00
333GPT-4.1$2.00
334o3$2.00
335o4 Mini Deep Research$2.00
336Perplexity: Sonar Deep Research$2.00
337Grok 4.5$2.00
338Grok 4.20$2.00
339Grok 4.20 0309 (Reasoning)$2.00
340Qwen3.7 Max$2.50
341Amazon: Nova Premier 1.0$2.50
342Cohere: Command R+ (08-2024)$2.50
343Inflection: Inflection 3 Pi$2.50
344Inflection: Inflection 3 Productivity$2.50
345GPT Audio$2.50
346GPT-4o (2024-08-06)$2.50
347GPT-4o (2024-11-20)$2.50
348GPT-4o Audio$2.50
349GPT-4o Search Preview$2.50
350GPT-5 Image Mini$2.50
351GPT-5.4$2.50
352GPT-5.6 Terra (high)$2.50
353GPT-5.6 Terra (low)$2.50
354GPT-5.6 Terra (max)$2.50
355GPT-5.6 Terra (medium)$2.50
356GPT-5.6 Terra (Non-reasoning)$2.50
357GPT-5.6 Terra (xhigh)$2.50
358OpenAI: GPT-4o$2.50
359OpenAI: GPT-5.6 Terra Pro$2.50
360Llama 3.1 Instruct 405B$2.75
361Mistral Medium$2.75
362Claude 3 Sonnet$3.00
363Claude 3.5 Sonnet (June '24)$3.00
364Claude 3.5 Sonnet (Oct '24)$3.00
365Claude 3.7 Sonnet$3.00
366Claude 4 Sonnet (Reasoning)$3.00
367Claude 4.5 Sonnet (Non-reasoning)$3.00
368Claude 4.5 Sonnet (Reasoning)$3.00
369Claude Sonnet 4$3.00
370Claude Sonnet 4.5$3.00
371Claude Sonnet 4.6$3.00
372Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)$3.00
373Claude Sonnet 4.6 (Non-reasoning, Low Effort)$3.00
374Command-R+ (Apr '24)$3.00
375Kimi K3$3.00
376Magnum v4 72B$3.00
377OpenAI: GPT-3.5 Turbo 16k$3.00
378Perplexity: Sonar Pro Search$3.00
379Sao10K: Llama 3.1 70B Hanami x1$3.00
380Grok 3 Beta$3.00
381Goliath 120B$3.75
382AionLabs: Aion-1.0$4.00
383Mistral Large 3$4.00
384Grok 3$4.00
385Claude Opus 4.5$5.00
386Claude Opus 4.5 (Reasoning)$5.00
387Claude Opus 4.6$5.00
388Claude Opus 4.6 (Adaptive Reasoning, Max Effort)$5.00
389Claude Opus 4.7$5.00
390Claude Opus 4.8 (Adaptive Reasoning, Max Effort)$5.00
391GPT Chat Latest$5.00
392GPT-5.5$5.00
393GPT-5.5 Instant (June 2026)$5.00
394GPT-5.5 Instant (May 2026)$5.00
395GPT-5.6 Sol (high)$5.00
396GPT-5.6 Sol (low)$5.00
397GPT-5.6 Sol (max)$5.00
398GPT-5.6 Sol (medium)$5.00
399GPT-5.6 Sol (Non-reasoning)$5.00
400GPT-5.6 Sol (xhigh)$5.00
401OpenAI: GPT-4o (2024-05-13)$5.00
402OpenAI: GPT-5.6 Sol Pro$5.00
403Grok 4$5.50
404GPT-5.4 Image 2$8.00
405Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)$10.00
406GPT-4 Turbo$10.00
407GPT-4 Turbo Preview$10.00
408GPT-5 Image$10.00
409o3 Deep Research$10.00
410OpenAI: GPT-4 Turbo (older v1106)$10.00
411Claude 3 Opus$15.00
412Claude 4 Opus (Reasoning)$15.00
413Claude 4.1 Opus (Non-reasoning)$15.00
414Claude 4.1 Opus (Reasoning)$15.00
415Claude Opus 4$15.00
416Claude Opus 4.1$15.00
417GPT-5 Pro$15.00
418o1$15.00
419o1-preview$16.50
420o3 Pro$20.00
421GPT-5.2 Pro$21.00
422Claude Opus 4.6 (Fast)$30.00
423Claude Opus 4.7 (Fast)$30.00
424GPT-5.4 Pro$30.00
425OpenAI: GPT-4$30.00
426o1-pro$150.00

APIs with Free Tier

These models offer free API access (with rate limits). Ideal for prototypes and low-volume projects.

Apriel-v1.5-15B-Thinker

ServiceNow

Free

Apriel-v1.6-15B-Thinker

ServiceNow

Free

Arctic Instruct

Snowflake

Free

Claude 2.0

Anthropic

Free

Claude 2.1

Anthropic

Free

Claude 3.7 Sonnet (thinking)

Anthropic

Free

Claude Instant

Anthropic

Free

Command A+

Cohere

Free

DBRX Instruct

Databricks

Free

DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)

Nous Research

Free

DeepHermes 3 - Mistral 24B Preview (Non-reasoning)

Nous Research

Free

DeepSeek Coder V2 Lite Instruct

DeepSeek

Free

DeepSeek LLM 67B Chat (V1)

DeepSeek

Free

DeepSeek R1 0528 Qwen3 8B

DeepSeek

Free

DeepSeek R1 Distill Llama 8B

DeepSeek

Free

DeepSeek R1 Distill Qwen 1.5B

DeepSeek

Free

DeepSeek R1 Distill Qwen 14B

DeepSeek

Free

DeepSeek V3.2 Speciale

DeepSeek

Free

DeepSeek-Coder-V2

DeepSeek

Free

DeepSeek-V2-Chat

DeepSeek

Free

DeepSeek-V2.5

DeepSeek

Free

DeepSeek-V2.5 (Dec '24)

DeepSeek

Free

DeepSeek: R1 Distill Qwen 32B

DeepSeek

Free

Devstral 2

Mistral

Free

Devstral Small (May '25)

Mistral

Free

Devstral Small 2

Mistral

Free

DiffusionGemma 26B A4B

Google

Free

Doubao Seed Code

ByteDance Seed

Free

ERNIE 5.0 Thinking Preview

Baidu

Free

Exaone 4.0 1.2B (Non-reasoning)

LG AI Research

Free

EXAONE 4.0 32B (Non-reasoning)

LG AI Research

Free

EXAONE 4.0 32B (Reasoning)

LG AI Research

Free

EXAONE 4.5 33B

LG AI

Free

Falcon-H1R-7B

TII UAE

Free

Gemini 1.0 Pro

Google

Free

Gemini 1.0 Ultra

Google

Free

Gemini 1.5 Flash (May '24)

Google

Free

Gemini 1.5 Flash (Sep '24)

Google

Free

Gemini 1.5 Flash-8B

Google

Free

Gemini 1.5 Pro (May '24)

Google

Free

Gemini 1.5 Pro (Sep '24)

Google

Free

Gemini 2.0 Flash (experimental)

Google

Free

Gemini 2.0 Flash Thinking Experimental (Dec '24)

Google

Free

Gemini 2.0 Flash Thinking Experimental (Jan '25)

Google

Free

Gemini 2.0 Flash-Lite (Feb '25)

Google

Free

Gemini 2.0 Flash-Lite (Preview)

Google

Free

Gemini 2.0 Pro Experimental (Feb '25)

Google

Free

Gemini 2.5 Flash Preview (Non-reasoning)

Google

Free

Gemini 2.5 Flash Preview (Sep '25) (Reasoning)

Google

Free

Gemini 2.5 Pro Preview (Mar' 25)

Google

Free

Gemini 3 Deep Think

Google

Free

Gemma 3 12B

Google

Free

Gemma 3 1B Instruct

Google

Free

Gemma 3 270M

Google

Free

Gemma 3 27B

Google

Free

Gemma 3 4B

Google

Free

Gemma 3n E2B Instruct

Google

Free

Gemma 3n E4B Instruct Preview (May '25)

Google

Free

Gemma 4 E2B (Non-reasoning)

Google

Free

Gemma 4 E2B (Reasoning)

Google

Free

Gemma 4 E4B (Non-reasoning)

Google

Free

Gemma 4 E4B (Reasoning)

Google

Free

GLM 5V Turbo (Reasoning)

Z.ai

Free

GLM-4.5 (Reasoning)

Z.ai

Free

GLM-5-Turbo

Z.ai

Free

GPT-3.5 Turbo (0613)

OpenAI

Free

GPT-4.5 (Preview)

OpenAI

Free

GPT-4o (ChatGPT)

OpenAI

Free

GPT-4o (March 2025, chatgpt-4o-latest)

OpenAI

Free

GPT-4o mini Realtime (Dec '24)

OpenAI

Free

GPT-4o Realtime (Dec '24)

OpenAI

Free

GPT-5.5 Pro

OpenAI

Free

Granite 4.0 1B

IBM

Free

Granite 4.0 350M

IBM

Free

Granite 4.0 H 1B

IBM

Free

Granite 4.0 H 350M

IBM

Free

Granite 4.0 Micro

IBM

Free

Granite 4.1 30B

IBM

Free

Granite 4.1 3B

IBM

Free

Grok 2 (Dec '24)

xAI

Free

Grok 4.1 Fast

xAI

Free

Grok Beta

xAI

Free

Grok Code Fast 1

xAI

Free

Grok-1

xAI

Free

HyperCLOVA X SEED Think (32B)

Naver

Free

INTELLECT-3

Prime Intellect

Free

Jamba 1.7 Mini

AI21 Labs

Free

Jamba Reasoning 3B

AI21 Labs

Free

JT-35B-Flash

China Mobile

Free

JT-4.1 Flash 236B A21B

China Mobile

Free

JT-MINI

China Mobile

Free

K-EXAONE (Reasoning)

LG AI

Free

K2 Think V2

MBZUAI Institute of Foundation Models

Free

K2-V2 (high)

MBZUAI Institute of Foundation Models

Free

K2-V2 (medium)

MBZUAI Institute of Foundation Models

Free

KAT-Coder-Pro V1

KwaiKAT

Free

Kimi Linear 48B A3B Instruct

Kimi

Free

LFM 40B

Liquid AI

Free

LFM2 1.2B

Liquid AI

Free

LFM2 2.6B

Liquid AI

Free

LFM2 8B A1B

Liquid AI

Free

LFM2-24B-A2B

LiquidAI

Free

LFM2.5-1.2B-Instruct

Liquid AI

Free

LFM2.5-1.2B-Thinking

Liquid AI

Free

LFM2.5-8B-A1B

Liquid AI

Free

LFM2.5-VL-1.6B

Liquid AI

Free

Ling-1T

InclusionAI

Free

Ling-mini-2.0

InclusionAI

Free

Llama 2 Chat 13B

Meta

Free

Llama 2 Chat 70B

Meta

Free

Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)

NVIDIA

Free

Llama 3.1 Tulu3 405B

Allen Institute for AI

Free

Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)

NVIDIA

Free

Llama 3.3 Nemotron Super 49B v1 (Reasoning)

NVIDIA

Free

Llama 65B

Meta

Free

LongCat 2.0

LongCat

Free

LongCat Flash Lite

LongCat

Free

Magistral Medium 1

Mistral

Free

Magistral Small 1

Mistral

Free

Mi:dm K 2.5 Pro

Korea Telecom

Free

Mi:dm K 2.5 Pro Preview

Korea Telecom

Free

MiMo-V2-Flash (Feb 2026)

Xiaomi

Free

MiMo-V2-Omni-0327

Xiaomi

Free

MiniCPM-V 4.6 1.3B

OpenBMB

Free

MiniCPM5-1B (Non-reasoning)

OpenBMB

Free

MiniMax M1 40k

MiniMax

Free

Mistral: Saba

Mistral AI

Free

Mixtral 8x22B Instruct

Mistral

Free

Molmo 7B-D

Allen Institute for AI

Free

Molmo2-8B

Allen Institute for AI

Free

Motif-2-12.7B-Reasoning

Motif Technologies

Free

Muse Spark

Meta

Free

Nanbeige4.1-3B

Nanbeige

Free

Nemotron Cascade 2 30B A3B

NVIDIA

Free

North Mini Code

Cohere

Free

NVIDIA Nemotron 3 Nano 4B

NVIDIA

Free

o1-mini

OpenAI

Free

OLMo 2 32B

Allen Institute for AI

Free

OLMo 2 7B

Allen Institute for AI

Free

Olmo 3 32B Think

AllenAI

Free

Olmo 3 7B Think

Allen Institute for AI

Free

Olmo 3.1 32B Instruct

AllenAI

Free

Olmo 3.1 32B Think

Allen Institute for AI

Free

OpenChat 3.5 (1210)

OpenChat

Free

PALM-2

Google

Free

Phi-3 Mini Instruct 3.8B

Microsoft

Free

Phi-4 Mini Instruct

Microsoft

Free

Phi-4 Multimodal Instruct

Microsoft

Free

Qwen Chat 14B

Alibaba

Free

Qwen Chat 72B

Alibaba

Free

Qwen1.5 Chat 110B

Alibaba

Free

Qwen2 Instruct 72B

Alibaba

Free

Qwen2.5 Coder 32B Instruct

Alibaba

Free

Qwen2.5 Coder Instruct 7B

Alibaba

Free

Qwen2.5 Instruct 32B

Alibaba

Free

Qwen2.5 Max

Alibaba

Free

Qwen3 0.6B (Non-reasoning)

Alibaba

Free

Qwen3 0.6B (Reasoning)

Alibaba

Free

Qwen3 1.7B (Non-reasoning)

Alibaba

Free

Qwen3 1.7B (Reasoning)

Alibaba

Free

Qwen3 4B (Non-reasoning)

Alibaba

Free

Qwen3 4B (Reasoning)

Alibaba

Free

Qwen3 4B 2507 (Reasoning)

Alibaba

Free

Qwen3 4B 2507 Instruct

Alibaba

Free

Qwen3 VL 4B (Reasoning)

Alibaba

Free

Qwen3 VL 4B Instruct

Alibaba

Free

Qwen3.5 0.8B (Non-reasoning)

Alibaba

Free

Qwen3.5 0.8B (Reasoning)

Alibaba

Free

Qwen3.5 2B (Reasoning)

Alibaba

Free

Qwen3.5 9B (Reasoning)

Alibaba

Free

QwQ 32B-Preview

Alibaba

Free

R1 1776

Perplexity

Free

Ring-1T

InclusionAI

Free

Sarvam M (Reasoning)

Sarvam

Free

Solar Open 100B (Reasoning)

Upstage

Free

Solar Pro 2 (Non-reasoning)

Upstage

Free

Solar Pro 2 (Preview) (Non-reasoning)

Upstage

Free

Solar Pro 2 (Preview) (Reasoning)

Upstage

Free

Solar Pro 3

Upstage

Free

Sonar

Perplexity

Free

Sonar Reasoning

Perplexity

Free

Sonar Reasoning Pro

Perplexity

Free

Step3 VL 10B

StepFun

Free

Tiny Aya Global

Cohere

Free

Tri-21B-Think

Trillion Labs

Free

Tri-21B-think Preview

Trillion Labs

Free

Xiaomi: MiMo-V2-Omni

Xiaomi

Free

Xiaomi: MiMo-V2-Pro

Xiaomi

Free

AI API Pricing Guide

How Token-Based Pricing Works

Most LLM APIs charge per token processed, split into two categories: input tokens (the text you send — your prompt, context, and history) and outputtokens (the model's generated response). Output pricing is typically 2–4× higher than input, as generation requires more compute.

In English, 1,000 tokens correspond to roughly 750 words. A full A4 page of text contains between 600 and 900 tokens.

Real-World Monthly Cost Example

Consider a company using the GPT-4o API to process 100 emails per day, with an average prompt of 800 tokens and a 300-token response. That's 110,000 tokens/day × 30 days = 3.3 million tokens/month. At $2.50/M input tokens and $10/M output:

  • Input: 2.4M tokens × $2.50/M = $6.00/mo
  • Output: 0.9M tokens × $10/M = $9.00/mo
  • Total: $15.00/mo

The same volume with Claude Haiku (~$0.25/M input) would cost only ~$1.73/mo — a significant saving when maximum quality isn't critical.

Strategies to Reduce API Costs

1. Pick the right model for each task: simple text classification can use Gemini Flash or Claude Haiku; reserve GPT-4o or Claude Opus for tasks that truly need advanced reasoning.

2. Compress your prompts: avoid repeating unnecessary context. Well-implemented RAG systems send only the relevant passages, not the entire document.

3. Cache responses: if the same prompt is sent repeatedly (e.g., product categorization), store results and reuse them. Providers like Anthropic offer prompt caching at a discount.

4. Use open-source models via third-party APIs: Groq, Together AI, and Fireworks serve models like Llama and Qwen at $0.01–$0.20/M tokens — 10–100× cheaper than proprietary frontier models.

Frequently Asked Questions about API Costs

How much does the GPT-4o API cost per token?

The GPT-4o API costs $2.50 per million input tokens and $10.00 per million output tokens (2026 pricing). For a business sending 1 million tokens per day, the monthly cost would be approximately $75 for input alone. Output tokens are 4× more expensive, so optimizing prompt length has a significant impact on cost.

What is the cheapest AI API available?

Open source models like Qwen, Llama and Gemma can be accessed via third-party APIs (Groq, Together AI, Fireworks) for fractions of a cent per million tokens — as low as $0.01–$0.10/M tokens. Among proprietary APIs, Gemini Flash and Claude Haiku are the most affordable at $0.08–$0.25/M input tokens.

What are tokens and how do I estimate my project cost?

Tokens are text units that LLMs process — in English, 1 token ≈ 4 characters. A standard A4 page has ~600–800 tokens. To estimate cost: (input tokens + output tokens) × price/1M tokens. Example: 500-token prompt + 300-token response = 800 tokens × model price.

Should I use the API or subscribe to ChatGPT Plus/Claude Pro?

For moderate personal use, a subscription ($20/month) is usually more economical. For heavy usage or product integration, the API is more flexible and scalable. The breakeven point typically occurs when your API token consumption exceeds the equivalent value of the monthly subscription.

How have AI API prices changed over time?

Prices have dropped dramatically: GPT-4 cost $30/M tokens in 2023; equivalent models now cost $2–5/M. The trend is continuous decline as competition increases. We update this table weekly — always verify official pricing before committing your budget.

Explore the Benchmark

Overall RankingCompare ModelsOpen Source ModelsMethodologyFree ToolsWhich AI to use? →