Presyo ng LLM API ayon sa model
Na-update
Presyo ng 428 na model mula sa 41 na provider, 1,671 na presyo lahat, bawat isa binasa sa sariling pricing page ng provider, may source link at petsa (pinakabago: Nobyembre 15, 2026). Nasa bawat page ng model ang input, cached input at output price ng bawat provider, ang sariling presyo ng gumawa, mga service tier at ang buwanang gastos ng tatlong totoong workload.
Naka-presyo ang LLM API kada milyong token, na may hiwalay na presyo para sa input, cached input at output. Para sa mga open-weight model gaya ng DeepSeek, Qwen, GLM, Kimi at Llama, maraming provider ang nagse-serve ng parehong weights sa presyong malaki ang pagkakaiba. May presyo sa sariling API ng gumawa ang 160 sa 428 na model. Nasa table sa ibaba ang 198 na model na may standard na presyo mula sa 2 o higit pang provider. Iisa lang ang provider ng natitirang 230; nasa maikling listahan ng mga model na iisa ang provider (Ingles) sila, at hindi isinasama sa search results ang mga page nila hanggang may pangalawang provider na maglista sa kanila.
Hinahanap ang pinakamababang presyo? Ang pinakamurang LLM API ngayon: nira-rank ang mga model sa iisang blended na presyo, sa kabuuan at ayon sa laki. Log ng pagbabago ng presyo (Ingles): bawat presyong gumalaw sa pagitan ng mga pagsuri namin. Presyo ayon sa provider (Ingles): ang buong price list at mga tier ng bawat provider; sa presyo ayon sa gumawa ng model, naka-grupo ang mga model ayon sa kompanyang gumawa sa kanila. Mga open-weight model ayon sa use case (Ingles): inaayos ang mga model ayon sa nailathalang benchmark score, iisang source at setting kada table, katabi ng mga presyong ito.
| Model | Gumawa | Mga provider | Pinakamurang input | Pinakamurang output | Pinakamurang blended 3:1 | Median in / out | Presyo ng gumawa in / out | Max context |
|---|---|---|---|---|---|---|---|---|
| Qwen · 35 na model | ||||||||
| Qwen 2.5 7B (Ingles) | Qwen | 3 | US$0.05 | US$0.05 | US$0.05 | US$0.05 / US$0.13 | — | 32,768 |
| Qwen3-8B (Ingles) | Qwen | 2 | US$0.06 | US$0.06 | US$0.06 | US$0.0885 / US$0.2575 | — | 131,072 |
| Qwen3.5 9B (Ingles) | Qwen | 4 | US$0.10 | US$0.15 | US$0.1125 | US$0.135 / US$0.20 | — | 262,144 |
| Qwen3 30B A3B (fp8) (Ingles) | Qwen | 3 | US$0.051 | US$0.25 | US$0.122 | US$0.08 / US$0.335 | — | 40,960 |
| Qwen3 Coder 30B A3B (Ingles) | Qwen | 3 | US$0.07 | US$0.28 | US$0.1225 | US$0.125 / US$0.45 | — | 262,144 |
| Qwen3-14B (Ingles) | Qwen | 3 | US$0.07 | US$0.24 | US$0.1225 | US$0.12 / US$0.24 | — | 131,072 |
| Qwen3 32B FP8 (Ingles) | Qwen | 5 | US$0.08 | US$0.28 | US$0.13 | US$0.102 / US$0.57 | — | 131,072 |
| Qwen3-30B-A3B-Instruct-2507 (Ingles) | Qwen | 2 | US$0.09 | US$0.30 | US$0.1425 | US$0.1075 / US$0.375 | — | 262,144 |
| Qwen3 235B A22B Instruct 2507 FP8 (Ingles) | Qwen | 5 | US$0.0875 | US$0.35 | US$0.1531 | US$0.14 / US$0.75 | — | 262,144 |
| Qwen3.8 Flash (Ingles) | Qwen | 5 | US$0.113 | US$0.382 | US$0.1803 | US$0.15 / US$0.47 | US$0.15 / US$0.47 | 1,048,575 |
| Qwen3 VL 30B A3B Instruct (Ingles) | Qwen | 4 | US$0.15 | US$0.50 | US$0.2375 | US$0.175 / US$0.65 | — | 262,144 |
| Qwen3 Coder Next (Ingles) | Qwen | 3 | US$0.12 | US$0.80 | US$0.29 | US$0.16 / US$1.15 | — | 32,768 |
| Qwen3.6 35B A3B | Qwen | 10 | US$0.10 | US$0.89 | US$0.2975 | US$0.168 / US$1.0569 | US$0.375 / US$2.25 | 262,144 |
| Qwen3-VL 8B (Ingles) | Qwen | 2 | US$0.18 | US$0.68 | US$0.305 | US$0.215 / US$0.715 | — | 262,144 |
| Qwen3.8 27B · nasa BroadNet Inference din | Qwen | 15 | US$0.039 | US$1.25 | US$0.3417 | US$0.40 / US$3 | US$0.50 / US$3 | 1,000,000 |
| Qwen3 Next 80B A3B Instruct (Ingles) | Qwen | 5 | US$0.09 | US$1.10 | US$0.3425 | US$0.125 / US$1.15 | — | 262,144 |
| Qwen3.5 35B A3B | Qwen | 7 | US$0.14 | US$1 | US$0.355 | US$0.24 / US$1.80 | US$0.25 / US$2 | 262,144 |
| Qwen2.5-72B-Instruct (Ingles) | Qwen | 3 | US$0.36 | US$0.40 | US$0.37 | US$0.37 / US$0.40 | — | 32,768 |
| Qwen3 VL 235B A22B (Ingles) | Qwen | 3 | US$0.20 | US$0.88 | US$0.37 | US$0.21 / US$1.90 | — | 262,144 |
| Qwen-MT Plus (Ingles) | Qwen | 2 | US$0.25 | US$0.75 | US$0.375 | US$1.355 / US$4.06 | US$2.46 / US$7.37 | 16,384 |
| Qwen3 235B A22B FP8 (Ingles) | Qwen | 2 | US$0.17 | US$1.09 | US$0.40 | US$0.3125 / US$1.455 | — | 32,768 |
| Qwen3 Next 80B A3B Thinking (Ingles) | Qwen | 2 | US$0.15 | US$1.20 | US$0.4125 | US$0.15 / US$1.35 | — | 262,144 |
| Qwen3 Coder 480B A35B | Qwen | 6 | US$0.22 | US$1 | US$0.4375 | US$0.325 / US$1.525 | — | 262,144 |
| Qwen3.7 Plus (Ingles) | Qwen | 2 | US$0.32 | US$1.28 | US$0.56 | US$0.36 / US$1.44 | US$0.40 / US$1.60 | — |
| Qwen3.5 27B | Qwen | 7 | US$0.20 | US$1.75 | US$0.5875 | US$0.30 / US$2.40 | US$0.30 / US$2.40 | 262,144 |
| Qwen3.6 27B | Qwen | 7 | US$0.25 | US$2 | US$0.6875 | US$0.32 / US$3.20 | US$0.60 / US$3.60 | 262,144 |
| Qwen3.5 122B A10B (Ingles) | Qwen | 4 | US$0.26 | US$2.08 | US$0.715 | US$0.40 / US$3.20 | US$0.40 / US$3.20 | 262,144 |
| Qwen2.5 Coder 32B Instruct (Ingles) | Qwen | 2 | US$0.66 | US$1 | US$0.745 | US$0.67 / US$1.10 | — | 32,768 |
| Qwen3.5 397B A17B | Qwen | 7 | US$0.45 | US$3 | US$1.087 | US$0.55 / US$3.60 | US$0.60 / US$3.60 | 262,144 |
| Qwen3.6 Plus (Ingles) | Qwen | 3 | US$0.50 | US$3 | US$1.125 | US$0.50 / US$3 | US$0.50 / US$3 | 262,144 |
| Qwen3 235B A22B Thinking 2507 FP8 (Ingles) | Qwen | 3 | US$0.35 | US$3.50 | US$1.212 | US$0.45 / US$3.50 | — | 131,072 |
| Qwen3.8 2.4T A95B | Qwen | 6 | US$1 | US$4 | US$1.75 | US$2 / US$6 | US$2 / US$6 | 1,048,576 |
| Qwen3.7 Max (Ingles) | Qwen | 5 | US$1.25 | US$3.75 | US$1.875 | US$2.50 / US$7.50 | US$2.50 / US$7.50 | 1,000,000 |
| Qwen3.8 Max | Qwen | 6 | US$1.65 | US$4.951 | US$2.475 | US$2 / US$6 | US$2 / US$6 | 1,000,000 |
| Qwen3.8 Max (0902) (Ingles) | Qwen | 3 | US$1.65 | US$4.951 | US$2.475 | US$2 / US$6 | US$2 / US$6 | 1,000,000 |
| OpenAI · 30 na model | ||||||||
| GPT OSS 20B | OpenAI | 10 | US$0.018 | US$0.09 | US$0.036 | US$0.04 / US$0.19 | — | 131,072 |
| OpenAI GPT OSS 120B | OpenAI | 21 | US$0.03 | US$0.17 | US$0.065 | US$0.10 / US$0.60 | — | 131,072 |
| GPT OSS Safeguard 20B (Ingles) | OpenAI | 3 | US$0.07 | US$0.20 | US$0.1025 | US$0.075 / US$0.30 | — | 131,072 |
| GPT-5 Nano (Ingles) | OpenAI | 3 | US$0.05 | US$0.40 | US$0.1375 | US$0.05 / US$0.40 | US$0.05 / US$0.40 | — |
| GPT-6 Luna | OpenAI | 6 | US$0.10 | US$0.50 | US$0.20 | US$0.10 / US$0.50 | US$0.10 / US$0.50 | 1,050,000 |
| GPT-4o mini (Ingles) | OpenAI | 3 | US$0.15 | US$0.60 | US$0.2625 | US$0.15 / US$0.60 | US$0.15 / US$0.60 | — |
| GPT-5.6 Luna | OpenAI | 6 | US$0.20 | US$1.20 | US$0.45 | US$0.20 / US$1.20 | US$0.20 / US$1.20 | 1,050,000 |
| GPT-5.4 Nano (Ingles) | OpenAI | 3 | US$0.20 | US$1.25 | US$0.4625 | US$0.20 / US$1.25 | US$0.20 / US$1.25 | — |
| GPT-5 Mini (Ingles) | OpenAI | 3 | US$0.25 | US$2 | US$0.6875 | US$0.25 / US$2 | US$0.25 / US$2 | — |
| GPT-4.1 Mini (Ingles) | OpenAI | 2 | US$0.40 | US$1.60 | US$0.70 | US$0.40 / US$1.60 | US$0.40 / US$1.60 | — |
| GPT-5.4 Mini (Ingles) | OpenAI | 3 | US$0.75 | US$4.50 | US$1.688 | US$0.75 / US$4.50 | US$0.75 / US$4.50 | — |
| GPT-6.1 Terra (Ingles) | OpenAI | 4 | US$1 | US$5 | US$2 | US$1 / US$5 | US$1 / US$5 | 1,050,000 |
| GPT-5 (2025-08-07) (Ingles) | OpenAI | 3 | US$1.25 | US$10 | US$3.438 | US$1.25 / US$10 | US$1.25 / US$10 | — |
| GPT-5.1 (Ingles) | OpenAI | 2 | US$1.25 | US$10 | US$3.438 | US$1.25 / US$10 | US$1.25 / US$10 | — |
| GPT-4.1 (2025-04-14) (Ingles) | OpenAI | 3 | US$2 | US$8 | US$3.50 | US$2 / US$8 | US$2 / US$8 | — |
| o3 (2025-04-16) (Ingles) | OpenAI | 3 | US$2 | US$8 | US$3.50 | US$2 / US$8 | US$2 / US$8 | — |
| GPT-6 Sol | OpenAI | 6 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,050,000 |
| GPT-6.1 Sol | OpenAI | 6 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,050,000 |
| GPT-4o (Ingles) | OpenAI | 2 | US$2.50 | US$10 | US$4.375 | US$2.50 / US$10 | US$2.50 / US$10 | — |
| GPT-5.6 Terra | OpenAI | 6 | US$2 | US$12 | US$4.50 | US$2 / US$12 | US$2 / US$12 | 1,050,000 |
| GPT-5.2 (Ingles) | OpenAI | 3 | US$1.75 | US$14 | US$4.813 | US$1.75 / US$14 | US$1.75 / US$14 | — |
| GPT-5.3 Codex (Ingles) | OpenAI | 3 | US$1.75 | US$14 | US$4.813 | US$1.75 / US$14 | US$1.75 / US$14 | — |
| GPT-5.4 (Ingles) | OpenAI | 4 | US$2.50 | US$15 | US$5.625 | US$2.50 / US$15 | US$2.50 / US$15 | 1,050,000 |
| GPT-5.6 Sol | OpenAI | 6 | US$4 | US$20 | US$8 | US$4 / US$20 | US$4 / US$20 | 1,050,000 |
| GPT Chat Latest (Ingles) | OpenAI | 2 | US$5 | US$30 | US$11.25 | US$5 / US$30 | US$5 / US$30 | — |
| GPT-5.5 | OpenAI | 6 | US$5 | US$30 | US$11.25 | US$5 / US$30 | US$5 / US$30 | 1,050,000 |
| GPT-6 Astra | OpenAI | 6 | US$10 | US$50 | US$20 | US$10 / US$50 | US$10 / US$50 | 1,050,000 |
| GPT-5 Pro (Ingles) | OpenAI | 2 | US$15 | US$120 | US$41.25 | US$15 / US$120 | US$15 / US$120 | — |
| GPT-5.2 Pro (Ingles) | OpenAI | 2 | US$21 | US$168 | US$57.75 | US$21 / US$168 | US$21 / US$168 | — |
| GPT-5.4 Pro (Ingles) | OpenAI | 3 | US$30 | US$180 | US$67.50 | US$30 / US$180 | US$30 / US$180 | — |
| DeepSeek · 20 na model | ||||||||
| DeepSeek V4 Flash 0731 | DeepSeek | 15 | US$0.06 | US$0.18 | US$0.09 | US$0.22 / US$0.66 | — | 1,048,576 |
| DeepSeek V4 Flash | DeepSeek | 8 | US$0.09 | US$0.18 | US$0.1125 | US$0.139 / US$0.28 | — | 1,048,576 |
| DeepSeek R1 Distill Qwen 7B (Ingles) | DeepSeek | 2 | US$0.10 | US$0.20 | US$0.125 | US$0.39 / US$0.50 | — | 131,072 |
| DeepSeek V4.1 Flash 1017 | DeepSeek | 16 | US$0.098 | US$0.294 | US$0.147 | US$0.30 / US$1.20 | US$0.30 / US$1.20 | 1,048,576 |
| DeepSeek R1 Distill Qwen 14B (Ingles) | DeepSeek | 2 | US$0.20 | US$0.20 | US$0.20 | US$0.3155 / US$0.66 | — | 131,072 |
| DeepSeek R1 Distill Llama 8B (Ingles) | DeepSeek | 2 | US$0.14 | US$0.32 | US$0.2025 | US$0.17 / US$0.355 | — | 131,072 |
| DeepSeek V4.1 Flash · nasa BroadNet Inference din | DeepSeek | 16 | US$0.13 | US$0.42 | US$0.21 | US$0.30 / US$1.20 | US$0.30 / US$1.20 | 1,048,576 |
| DeepSeek V3.2 | DeepSeek | 11 | US$0.2088 | US$0.3096 | US$0.234 | US$0.33 / US$0.48 | — | 163,840 |
| DeepSeek-V3.2-Exp (Ingles) | DeepSeek | 2 | US$0.27 | US$0.41 | US$0.305 | US$0.27 / US$0.41 | — | 163,840 |
| DeepSeek-V4-Flash-Vision-Exp (Ingles) | DeepSeek | 3 | US$0.2156 | US$0.6468 | US$0.3234 | US$0.44 / US$1.32 | — | 1,048,576 |
| DeepSeek R1 Distill Llama 70B (Ingles) | DeepSeek | 2 | US$0.25 | US$0.75 | US$0.375 | US$1.425 / US$1.875 | — | 131,072 |
| DeepSeek V3.1 | DeepSeek | 7 | US$0.25 | US$0.95 | US$0.425 | US$0.285 / US$1 | — | 163,840 |
| DeepSeek V3 (Ingles) | DeepSeek | 3 | US$0.25 | US$0.89 | US$0.4375 | US$0.32 / US$1 | — | 163,840 |
| DeepSeek-V3.1-Terminus (Ingles) | DeepSeek | 3 | US$0.27 | US$1 | US$0.4525 | US$0.27 / US$1 | — | 163,840 |
| DeepSeek V3 0324 (Ingles) | DeepSeek | 3 | US$0.29 | US$1.14 | US$0.5025 | US$0.385 / US$1.625 | — | 131,072 |
| DeepSeek R1 Distill Qwen 32B (Ingles) | DeepSeek | 3 | US$0.497 | US$0.90 | US$0.60 | US$0.50 / US$4.80 | — | 131,072 |
| DeepSeek R1 0528 (Ingles) | DeepSeek | 5 | US$0.385 | US$1.625 | US$0.695 | US$0.57 / US$2.29 | — | 163,840 |
| DeepSeek R1 (Ingles) | DeepSeek | 4 | US$0.50 | US$2.18 | US$0.92 | US$0.925 / US$3.79 | — | 163,840 |
| DeepSeek V4 Pro | DeepSeek | 11 | US$0.957 | US$1.914 | US$1.196 | US$1.65 / US$3.301 | US$1.32 / US$3.96 | 1,048,576 |
| DeepSeek V4 Pro 0813 | DeepSeek | 15 | US$0.85 | US$2.60 | US$1.30 | US$1.32 / US$3.96 | US$1.32 / US$3.96 | 1,048,576 |
| Google · 19 na model | ||||||||
| Gemma 3 4B (Ingles) | 3 | US$0.04 | US$0.08 | US$0.05 | US$0.05 / US$0.10 | — | 131,072 | |
| Gemma 3 12B (Ingles) | 4 | US$0.05 | US$0.15 | US$0.075 | US$0.07 / US$0.22 | — | 131,072 | |
| Gemma 3 27B | 6 | US$0.08 | US$0.16 | US$0.10 | US$0.1095 / US$0.25 | — | 198,000 | |
| Gemma 4 26B A4B · nasa BroadNet Inference din | 11 | US$0.07 | US$0.30 | US$0.1375 | US$0.13 / US$0.40 | — | 262,144 | |
| gemma-4-31B-it-turbo (Ingles) | 2 | US$0.09 | US$0.34 | US$0.1525 | US$0.105 / US$0.355 | — | 262,144 | |
| Gemma 4 31B | 12 | US$0.10 | US$0.34 | US$0.16 | US$0.14 / US$0.40 | — | 262,144 | |
| Gemini 2.5 Flash-Lite (Ingles) | 2 | US$0.10 | US$0.40 | US$0.175 | US$0.10 / US$0.40 | US$0.10 / US$0.40 | — | |
| Gemini 3.1 Flash-Lite (gemini-3.1-flash-lite) (Ingles) | 3 | US$0.25 | US$1.50 | US$0.5625 | US$0.25 / US$1.50 | US$0.25 / US$1.50 | 1,000,000 | |
| Gemini 2.5 Flash (Ingles) | 3 | US$0.30 | US$2.50 | US$0.85 | US$0.30 / US$2.50 | US$0.30 / US$2.50 | 1,000,000 | |
| Gemini 3.5 Flash-Lite (Ingles) | 3 | US$0.30 | US$2.50 | US$0.85 | US$0.30 / US$2.50 | US$0.30 / US$2.50 | 1,048,576 | |
| Gemini 3.6 Flash (Ingles) | 3 | US$0.75 | US$3.75 | US$1.50 | US$0.75 / US$3.75 | US$0.75 / US$3.75 | 1,048,576 | |
| Gemini 3.7 Flash (Ingles) | 3 | US$0.75 | US$3.75 | US$1.50 | US$0.75 / US$3.75 | — | 1,048,576 | |
| Gemini 3.8 Flash (Ingles) | 3 | US$0.75 | US$3.75 | US$1.50 | US$0.75 / US$3.75 | US$0.75 / US$3.75 | 1,048,576 | |
| Gemini 4 Hydrogen (Ingles) | 3 | US$1.50 | US$7.50 | US$3 | US$1.50 / US$7.50 | US$1.50 / US$7.50 | — | |
| Gemini 3.5 Flash (Ingles) | 3 | US$1.50 | US$9 | US$3.375 | US$1.50 / US$9 | — | 1,048,576 | |
| Gemini 2.5 Pro (Ingles) | 3 | US$1.25 | US$10 | US$3.438 | US$1.25 / US$10 | US$1.25 / US$10 | 1,000,000 | |
| Gemini 4 Argon (Ingles) | 4 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,048,576 | |
| Gemini 4 Carbon (Ingles) | 4 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,048,576 | |
| Gemini 3.1 Pro Preview (Ingles) | 3 | US$2 | US$12 | US$4.50 | US$2 / US$12 | US$2 / US$12 | 1,048,576 | |
| Anthropic · 15 na model | ||||||||
| Claude Haiku 5.5 (Ingles) | Anthropic | 4 | US$0.10 | US$0.50 | US$0.20 | US$0.10 / US$0.50 | US$0.10 / US$0.50 | — |
| Claude Haiku 4.5 | Anthropic | 5 | US$1 | US$5 | US$2 | US$1 / US$5 | US$1 / US$5 | 200,000 |
| Claude Sonnet 5 (Ingles) | Anthropic | 5 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,000,000 |
| Claude Sonnet 5.5 | Anthropic | 6 | US$2 | US$10 | US$4 | US$2 / US$10 | US$2 / US$10 | 1,000,000 |
| Claude Sonnet 4.5 (Ingles) | Anthropic | 4 | US$3 | US$15 | US$6 | US$3 / US$15 | US$3 / US$15 | — |
| Claude Sonnet 4.6 (Ingles) | Anthropic | 5 | US$3 | US$15 | US$6 | US$3 / US$15 | US$3 / US$15 | 1,000,000 |
| Claude Opus 5.5 | Anthropic | 6 | US$4 | US$20 | US$8 | US$4 / US$20 | US$4 / US$20 | 1,000,000 |
| Claude Opus 4.5 (Ingles) | Anthropic | 4 | US$5 | US$25 | US$10 | US$5 / US$25 | US$5 / US$25 | — |
| Claude Opus 4.6 (Ingles) | Anthropic | 4 | US$5 | US$25 | US$10 | US$5 / US$25 | US$5 / US$25 | 1,000,000 |
| Claude Opus 4.7 (Ingles) | Anthropic | 5 | US$5 | US$25 | US$10 | US$5 / US$25 | US$5 / US$25 | 1,000,000 |
| Claude Opus 4.8 (Ingles) | Anthropic | 5 | US$5 | US$25 | US$10 | US$5 / US$25 | US$5 / US$25 | 1,000,000 |
| Claude Opus 5 (Ingles) | Anthropic | 5 | US$5 | US$25 | US$10 | US$5 / US$25 | US$5 / US$25 | 1,000,000 |
| Claude Fable 5.2 (Ingles) | Anthropic | 4 | US$8 | US$40 | US$16 | US$8 / US$40 | US$8 / US$40 | 1,000,000 |
| Claude Fable 5 | Anthropic | 5 | US$10 | US$50 | US$20 | US$10 / US$50 | US$10 / US$50 | 1,000,000 |
| Claude Fable 5.1 (Ingles) | Anthropic | 4 | US$10 | US$50 | US$20 | US$10 / US$50 | US$10 / US$50 | 1,000,000 |
| Z.ai · 13 na model | ||||||||
| GLM 5.3 Flash | Z.ai | 20 | US$0.075 | US$0.25 | US$0.1187 | US$0.15 / US$0.50 | US$0.15 / US$0.50 | 1,048,576 |
| GLM 4.7 Flash | Z.ai | 6 | US$0.06 | US$0.40 | US$0.145 | US$0.07 / US$0.40 | libre | 202,752 |
| GLM 4.5 Air (Ingles) | Z.ai | 4 | US$0.13 | US$0.85 | US$0.31 | US$0.17 / US$0.98 | US$0.20 / US$1.10 | 131,076 |
| GLM 4.6V (Ingles) | Z.ai | 2 | US$0.30 | US$0.90 | US$0.45 | US$0.30 / US$0.90 | US$0.30 / US$0.90 | 131,072 |
| GLM 5.3 FlashX (Ingles) | Z.ai | 2 | US$0.37 | US$1.25 | US$0.59 | US$0.37 / US$1.25 | US$0.37 / US$1.25 | 1,048,576 |
| GLM 4.7 | Z.ai | 8 | US$0.40 | US$1.75 | US$0.7375 | US$0.60 / US$2.20 | US$0.60 / US$2.20 | 204,800 |
| GLM 4.6 | Z.ai | 6 | US$0.43 | US$1.75 | US$0.76 | US$0.55 / US$2.10 | US$0.60 / US$2.20 | 204,800 |
| GLM 5.2 | Z.ai | 19 | US$0.5625 | US$1.80 | US$0.8719 | US$1.40 / US$4.40 | US$1.40 / US$4.40 | 1,048,576 |
| GLM 4.5V (Ingles) | Z.ai | 2 | US$0.60 | US$1.80 | US$0.90 | US$0.60 / US$1.80 | US$0.60 / US$1.80 | 65,536 |
| GLM 5 | Z.ai | 10 | US$0.60 | US$1.92 | US$0.93 | US$1 / US$3.20 | US$1 / US$3.20 | 204,800 |
| GLM-4.5-FP8 (Ingles) | Z.ai | 2 | US$0.60 | US$2.20 | US$1 | US$0.60 / US$2.20 | US$0.60 / US$2.20 | 131,072 |
| GLM 5.3 · nasa BroadNet Inference din | Z.ai | 24 | US$0.19 | US$2.20 | US$1.047 | US$1.40 / US$4.40 | US$1.40 / US$4.40 | 1,048,576 |
| GLM 5.1 | Z.ai | 12 | US$0.90 | US$2.90 | US$1.40 | US$1.40 / US$4.40 | US$1.40 / US$4.40 | 204,800 |
| Meta · 12 na model | ||||||||
| Llama 3.1 8B Instruct | Meta | 6 | US$0.02 | US$0.04 | US$0.025 | US$0.21 / US$0.22 | — | 131,072 |
| Llama 3.2 1B Instruct (Ingles) | Meta | 3 | US$0.027 | US$0.10 | US$0.0705 | US$0.027 / US$0.201 | — | 32,768 |
| Llama 3 8B Instruct (Ingles) | Meta | 3 | US$0.0925 | US$0.095 | US$0.09312 | US$0.282 / US$0.60 | — | 8,192 |
| Llama 3.2 3B Instruct (Ingles) | Meta | 4 | US$0.0509 | US$0.15 | US$0.1219 | US$0.1005 / US$0.335 | — | 128,000 |
| Llama 4 Scout 17B 16E Instruct | Meta | 6 | US$0.08 | US$0.30 | US$0.15 | US$0.175 / US$0.625 | — | 1,048,576 |
| Llama 3.3 70B | Meta | 14 | US$0.10 | US$0.32 | US$0.155 | US$0.625 / US$0.72 | — | 131,072 |
| Llama 3.2 11B Vision Instruct (Ingles) | Meta | 3 | US$0.049 | US$0.16 | US$0.16 | US$0.16 / US$0.371 | — | — |
| Llama 4 Maverick 17B 128E Instruct FP8 | Meta | 7 | US$0.24 | US$0.80 | US$0.3875 | US$0.27 / US$1 | — | 1,048,576 |
| Muse Glimmer 30B (Ingles) | Meta | 3 | US$0.30 | US$1.10 | US$0.50 | US$0.30 / US$1.20 | — | 131,072 |
| Llama 3.1 Instruct (70B) (Ingles) | Meta | 3 | US$0.293 | US$0.72 | US$0.72 | US$0.72 / US$2.253 | — | 32,768 |
| Muse Spark Open (Ingles) | Meta | 4 | US$1.25 | US$4.25 | US$2 | US$1.25 / US$4.25 | — | 1,048,576 |
| Llama 3 Instruct (70B) (Ingles) | Meta | 2 | US$2.65 | US$3.50 | US$2.862 | US$3.075 / US$5.90 | — | 8,192 |
| Mistral AI · 9 na model | ||||||||
| Mistral-Nemo-Instruct-2407 (Ingles) | Mistral AI | 2 | US$0.029 | US$0.03 | US$0.02925 | US$0.12325 / US$0.13875 | — | 131,072 |
| Mistral-Small-24B-Instruct-2501 (Ingles) | Mistral AI | 2 | US$0.05 | US$0.08 | US$0.0575 | US$0.1125 / US$0.185 | — | 32,768 |
| Ministral 3 3B (Ingles) | Mistral AI | 2 | US$0.10 | US$0.10 | US$0.10 | US$0.10 / US$0.10 | US$0.10 / US$0.10 | 262,144 |
| Mistral Small 3.2 24B Instruct (Ingles) | Mistral AI | 4 | US$0.075 | US$0.20 | US$0.1062 | US$0.091875 / US$0.275 | — | 256,000 |
| Ministral 3 8B (Ingles) | Mistral AI | 2 | US$0.15 | US$0.15 | US$0.15 | US$0.15 / US$0.15 | US$0.15 / US$0.15 | 262,144 |
| Mistral Small 3.1 24B Instruct (Ingles) | Mistral AI | 4 | US$0.10 | US$0.30 | US$0.15 | US$0.1614 / US$0.36025 | — | 32,768 |
| Ministral 3 14B (Ingles) | Mistral AI | 3 | US$0.20 | US$0.20 | US$0.20 | US$0.20 / US$0.20 | US$0.20 / US$0.20 | 262,144 |
| Mistral Large 3 (Ingles) | Mistral AI | 3 | US$0.50 | US$1.50 | US$0.75 | US$0.636 / US$1.908 | — | — |
| Mistral Medium 3 (Ingles) | Mistral AI | 2 | US$0.40 | US$2 | US$0.80 | US$1.79 / US$5.75 | — | — |
| Moonshot AI · 8 na model | ||||||||
| Kimi K2.6 | Moonshot AI | 13 | US$0.39 | US$1.80 | US$0.7425 | US$0.80 / US$3.50 | US$0.95 / US$4 | 262,144 |
| Kimi K2.5 | Moonshot AI | 8 | US$0.45 | US$2.25 | US$0.90 | US$0.60 / US$3 | — | 262,144 |
| Kimi K2 Thinking | Moonshot AI | 6 | US$0.60 | US$1.20 | US$0.90 | US$0.60 / US$2.50 | — | 262,144 |
| Kimi K2 Instruct (Ingles) | Moonshot AI | 2 | US$0.57 | US$2.30 | US$1.002 | US$0.585 / US$2.40 | — | 131,072 |
| Kimi K2 Instruct 0905 (Ingles) | Moonshot AI | 2 | US$0.60 | US$2.50 | US$1.075 | US$0.60 / US$2.50 | — | 262,144 |
| Kimi K2.7 Code | Moonshot AI | 12 | US$0.60 | US$3 | US$1.20 | US$0.90458 / US$3.90 | US$0.95 / US$4 | 262,144 |
| Kimi K3 | Moonshot AI | 18 | US$1.85 | US$9.50 | US$3.763 | US$3 / US$15 | US$3 / US$15 | 1,048,576 |
| Kimi K3.1 Pro | Moonshot AI | 13 | US$1.75 | US$9.50 | US$3.80 | US$2 / US$10 | US$2 / US$10 | 1,048,576 |
| MiniMax · 6 na model | ||||||||
| MiniMax M3.1 Flash | MiniMax | 9 | US$0.15 | US$0.60 | US$0.2625 | US$0.15 / US$0.60 | US$0.15 / US$0.60 | 1,048,576 |
| MiniMax M2.7 (Ingles) | MiniMax | 3 | US$0.21 | US$0.84 | US$0.3675 | US$0.30 / US$1.20 | — | 204,800 |
| MiniMax M3 | MiniMax | 10 | US$0.23 | US$0.96 | US$0.4125 | US$0.30 / US$1.20 | US$0.30 / US$1.20 | 1,048,576 |
| MiniMax M2.5 | MiniMax | 6 | US$0.27 | US$0.95 | US$0.44 | US$0.30 / US$1.20 | — | 204,800 |
| MiniMax M2 (Ingles) | MiniMax | 4 | US$0.2775 | US$1.11 | US$0.4856 | US$0.30 / US$1.20 | — | 204,800 |
| Minimax M2.1 (Ingles) | MiniMax | 3 | US$0.30 | US$1.20 | US$0.525 | US$0.30 / US$1.20 | — | 204,800 |
| xAI · 6 na model | ||||||||
| Grok 4.20 (non-reasoning) (Ingles) | xAI | 2 | US$1.25 | US$2.50 | US$1.563 | US$1.25 / US$2.50 | US$1.25 / US$2.50 | 1,000,000 |
| Grok 4.20 (reasoning) (Ingles) | xAI | 3 | US$1.25 | US$2.50 | US$1.563 | US$1.25 / US$2.50 | US$1.25 / US$2.50 | 1,000,000 |
| Grok 4.3 (Ingles) | xAI | 4 | US$1.25 | US$2.50 | US$1.563 | US$1.25 / US$2.50 | US$1.25 / US$2.50 | 1,000,000 |
| Grok 4.5 (Ingles) | xAI | 3 | US$2 | US$6 | US$3 | US$2 / US$6 | US$2 / US$6 | 2,000,000 |
| Grok 4.6 | xAI | 6 | US$2 | US$6 | US$3 | US$2 / US$6 | US$2 / US$6 | 2,000,000 |
| Grok 4.7 | xAI | 6 | US$2 | US$6 | US$3 | US$2 / US$6 | US$2 / US$6 | 524,288 |
| NVIDIA · 5 na model | ||||||||
| NVIDIA Nemotron 3.5 Lightning 30B A3B (Ingles) | NVIDIA | 5 | US$0.05 | US$0.16 | US$0.085 | US$0.06 / US$0.20 | — | 1,048,576 |
| NVIDIA Nemotron 3 Nano 30B A3B (Ingles) | NVIDIA | 3 | US$0.05 | US$0.20 | US$0.0875 | US$0.06 / US$0.24 | — | 262,144 |
| NVIDIA Nemotron 3 Super 120B A12B (Ingles) | NVIDIA | 3 | US$0.075 | US$0.40 | US$0.1563 | US$0.15 / US$0.65 | — | 262,144 |
| NVIDIA Nemotron 3.5 Super VL | NVIDIA | 6 | US$0.10 | US$0.40 | US$0.175 | US$0.175 / US$0.50 | — | 1,000,000 |
| NVIDIA Nemotron 3 Ultra | NVIDIA | 9 | US$0.50 | US$2.15 | US$0.9125 | US$0.60 / US$2.40 | — | 262,144 |
| StepFun · 3 na model | ||||||||
| Step-3.5-Flash (Ingles) | StepFun | 2 | US$0.10 | US$0.30 | US$0.15 | US$0.10 / US$0.30 | — | 262,144 |
| Step 3.7 Flash (Ingles) | StepFun | 3 | US$0.20 | US$1.15 | US$0.4375 | US$0.20 / US$1.15 | — | 262,144 |
| Step 5 Preview (Ingles) | StepFun | 4 | US$1 | US$2.70 | US$1.425 | US$1 / US$2.70 | — | 1,048,576 |
| Tencent · 3 na model | ||||||||
| Hy3 (Ingles) | Tencent | 4 | US$0.132 | US$0.528 | US$0.231 | US$0.14 / US$0.58 | — | 262,144 |
| Hy4 (Ingles) | Tencent | 3 | US$0.417 | US$1.2505 | US$0.6254 | US$0.417 / US$1.2505 | — | 1,048,576 |
| Hy4 Preview | Tencent | 6 | US$0.834 | US$2.501 | US$1.251 | US$0.834 / US$2.501 | — | 1,048,576 |
| Xiaomi · 3 na model | ||||||||
| MiMo V2.6 Flash (Ingles) | Xiaomi | 5 | US$0.112 | US$0.224 | US$0.14 | US$0.14 / US$0.28 | — | 1,050,000 |
| MiMo v2.5 (Ingles) | Xiaomi | 4 | US$0.14 | US$0.28 | US$0.175 | US$0.1445 / US$0.3865 | — | 1,048,576 |
| MiMo V2.6 Pro (Ingles) | Xiaomi | 3 | US$0.32625 | US$0.6525 | US$0.4078 | US$0.387 / US$0.783 | — | 1,050,000 |
| Cohere · 2 na model | ||||||||
| Command R (Ingles) | Cohere | 2 | US$0.15 | US$0.60 | US$0.2625 | US$0.325 / US$1.05 | US$0.15 / US$0.60 | 128,000 |
| Command A+ (Ingles) | Cohere | 2 | US$0.80 | US$3.20 | US$1.40 | — | libre | — |
| inclusionAI · 2 na model | ||||||||
| Ling 3.0 Flash (Ingles) | inclusionAI | 2 | US$0.06 | US$0.18 | US$0.09 | US$0.06 / US$0.18 | — | 262,144 |
| Ling 3.0 Flash VL (Ingles) | inclusionAI | 2 | US$0.06 | US$0.18 | US$0.09 | US$0.0675 / US$0.20 | — | 262,144 |
| Sao10K · 2 na model | ||||||||
| Llama 3 8B Stheno V3.2 (Ingles) | Sao10K | 2 | US$0.05 | US$0.05 | US$0.05 | US$0.21 / US$0.215 | — | 8,192 |
| L3.1-70B-Euryale-v2.2 (Ingles) | Sao10K | 2 | US$0.72 | US$0.72 | US$0.72 | US$0.785 / US$0.785 | — | 131,072 |
| IBM · 1 na model | ||||||||
| IBM Granite 4.2 8B (Ingles) | IBM | 2 | US$0.06 | US$0.15 | US$0.0825 | US$0.06 / US$0.20 | — | 131,072 |
| Microsoft · 1 na model | ||||||||
| Phi-4 (Ingles) | Microsoft | 3 | US$0.07 | US$0.14 | US$0.0875 | US$0.07 / US$0.14 | — | 131,072 |
| Reflection AI · 1 na model | ||||||||
| Beam 1.0 501B A23B (Ingles) | Reflection AI | 3 | US$0.40 | US$1.20 | US$0.60 | US$0.40 / US$1.20 | — | — |
| Ricursive AI · 1 na model | ||||||||
| Ripple-65B Preview (Ingles) | Ricursive AI | 2 | US$0.20 | US$0.60 | US$0.30 | US$0.20 / US$0.60 | — | 1,000,000 |
| Thinking Machines · 1 na model | ||||||||
| Inkling 2 Small (Ingles) | Thinking Machines | 3 | US$0.285 | US$0.95 | US$0.4512 | US$0.30 / US$1 | — | 1,048,576 |
198 na model na may 2+ provider; hiwalay na nakalista ang 230 na model na iisa ang provider (Ingles). USD kada 1M token, standard tier. Ang "pinakamura" ay ang pinakamababang standard na presyo para sa model sa mga provider na sinuri namin, hindi sa buong market; puwedeng magmula sa magkaibang provider ang minimum na input at output. Nasa bawat page ng model ang batch, flex, priority at fast tier, at ang mga libreng row, pero hindi sila kailanman nagtatakda ng "pinakamura". Binibilang ang mga promotional na presyo at may tag ang mga ito sa page ng model. Ang "Presyo ng gumawa" ay ang sariling API ng gumawa ng model, kung nagbebenta ito ng API (DeepSeek: presyo sa peak hours). 1,218 na standard na presyo lahat; pinakabagong pagsuri: Nobyembre 15, 2026. Papunta sa sariling mga page ng model ng BroadNet Inference ang mga link ng BroadNet. Sariling site ng BroadNet ang BroadRouter.
I-report ang maling presyo o numero (email sa info@broadnet.ai, nakalagay na ang link ng page na ito).
Paano ihambing ang presyo ng LLM API?
- I-presyo ang workload mo, hindi ang model. Mabigat sa input ang chatbot; mabigat sa output ang reasoning agent. Tinatanggap ng cost calculator (Ingles) ang traffic at cache rate mo.
- Tingnan ang cached input. Ipinapadala muli ng mga agent ang parehong prefix; puwedeng mas sulit ang murang cache price kaysa murang input price. Mga cached input token.
- Tingnan ang context, precision at tier. Puwedeng mas mababa ang context cap ng pinakamurang row, mag-serve ito ng FP4, o maging best-effort na tier. Ipinapakita ng bawat page ng model ang tatlong ito.
- Bilangin ang token kada model. Magkakaiba ang tokenizer, kaya magkaiba ang bilang ng token ng parehong prompt. Gamitin ang token counter (Ingles).
Ang parehong data bilang open dataset (CSV at JSON; CC BY 4.0 ang compilation namin, at kinikilala ang provider ng bawat presyo), may agwat ng presyo sa pagitan ng mga provider, presyo ng gumawa laban sa host at cached pricing kada model: ang LLM API price index (Ingles). Bawat indibidwal na presyo, may source URL, petsa at quote: i-download ang lahat ng presyo (Ingles).
Tingnan din ang paghahambing ng model laban sa model (Ingles) at ang presyo at plan ng BroadNet Inference.
Mga tanong
Aling LLM API ang pinakamura kada token?
Sa 198 na model na may dalawa o higit pang provider sa data namin (sariling pricing page ng 41 na provider, binasa Nobyembre 14, 2026 hanggang Nobyembre 15, 2026), ang Llama 3.1 8B Instruct ang may pinakamababang nakalistang standard na presyo: US$0.02 input at US$0.04 output kada 1M token sa DeepInfra. Gumagamit ang ranking ng pinakamurang LLM API ng nakasaad na quality filter, kaya puwedeng iba ang nangunguna roon. Bihirang ang pinakamurang model ang pinakaangkop sa isang task, kaya i-presyo ang workload mo.
Bakit iba-iba ang presyo ng parehong model?
Sine-serve ang mga open-weight model ng maraming kompanya sa magkakaibang hardware, precision (FP4, FP8, BF16) at service tier. May standard na presyo ang GLM 5.3 mula sa 24 na provider. Sa 197 na model na may dalawa o higit pang may-bayad na standard na presyo, ang pinakamahal ay median na 1.3× ng pinakamura sa 3:1 blend.
Saan galing ang mga presyong ito?
Mula sa sariling pricing page ng bawat provider. Binasa namin ang 1,671 na presyo mula sa 41 na provider sa tulong ng AI at ng page renderer, at iniimbak namin ang bawat isa kasama ang source URL nito, ang petsa ng pagsuri at isang quote mula sa page. Walang aggregator o router feed na ginagamit. Nasa methodology page ang buong patakaran. Galing sa sariling rate card ng BroadNet Inference ang mga row ng BroadNet at hindi naka-rank ang mga ito.
Ano ang ibig sabihin ng presyo kada token kung kada milyon ang presyo?
USD kada 1M token ang presyo ng mga provider. Hatiin sa 1,000,000 para makuha ang presyo ng isang token: ang US$0.30 kada 1M ay US$0.0000003 kada token. Magkahiwalay ang presyo ng input, cached input at output, at sinisingil bilang output ang mga reasoning token.
Paano ko tatantiyahin ang buwanang bill?
Gamitin ang LLM API cost calculator (Ingles): pumili ng workload, saka magdagdag ng alinman sa 428 na model sa pinakamurang standard na presyo nito. Bilangin muna ang totoong laki ng mga prompt mo gamit ang token counter (Ingles).
Basahin bilang MarkdownTanungin ang ChatGPTTanungin ang ClaudeTanungin ang Perplexity
Mga source
- AI21 Labs (sinuri noong Nobyembre 15, 2026)
- AkashML (sinuri noong Nobyembre 15, 2026)
- Alibaba Cloud Model Studio (sinuri noong Nobyembre 14, 2026)
- Amazon Bedrock (sinuri noong Nobyembre 14, 2026)
- Amazon Nova (Bedrock) (sinuri noong Nobyembre 14, 2026)
- Anthropic (sinuri noong Nobyembre 14, 2026)
- Baidu (sinuri noong Nobyembre 15, 2026)
- Baseten (sinuri noong Nobyembre 14, 2026)
- Cerebras (sinuri noong Nobyembre 14, 2026)
- Chutes (sinuri noong Nobyembre 15, 2026)
- Cloudflare Workers AI (sinuri noong Nobyembre 14, 2026)
- Cohere (sinuri noong Nobyembre 14, 2026)
- CoreWeave Forge (formerly W&B Inference) (sinuri noong Nobyembre 15, 2026)
- Crusoe Cloud (sinuri noong Nobyembre 15, 2026)
- DeepInfra (sinuri noong Nobyembre 14, 2026)
- DeepSeek (sinuri noong Nobyembre 14, 2026)
- DigitalOcean Gradient (Serverless Inference) (sinuri noong Nobyembre 14, 2026)
- Featherless (sinuri noong Nobyembre 15, 2026)
- Fireworks AI (sinuri noong Nobyembre 14, 2026)
- FriendliAI (sinuri noong Nobyembre 15, 2026)
- GMI Cloud (sinuri noong Nobyembre 15, 2026)
- Google (Gemini API) (sinuri noong Nobyembre 14, 2026)
- Google Cloud Vertex AI (sinuri noong Nobyembre 14, 2026)
- Groq (sinuri noong Nobyembre 14, 2026)
- Hyperbolic (sinuri noong Nobyembre 14, 2026)
- IBM watsonx.ai (sinuri noong Nobyembre 15, 2026)
- Microsoft Azure (Azure OpenAI + Foundry Models) (sinuri noong Nobyembre 14, 2026)
- MiniMax (sinuri noong Nobyembre 14, 2026)
- Mistral AI (sinuri noong Nobyembre 14, 2026)
- Moonshot AI (Kimi) (sinuri noong Nobyembre 14, 2026)
- Novita AI (sinuri noong Nobyembre 14, 2026)
- Nscale (sinuri noong Nobyembre 15, 2026)
- OpenAI (sinuri noong Nobyembre 14, 2026)
- Parasail (sinuri noong Nobyembre 14, 2026)
- Perplexity (sinuri noong Nobyembre 15, 2026)
- SambaNova Cloud (sinuri noong Nobyembre 14, 2026)
- SiliconFlow (sinuri noong Nobyembre 14, 2026)
- Together AI (sinuri noong Nobyembre 14, 2026)
- Venice (sinuri noong Nobyembre 15, 2026)
- xAI (sinuri noong Nobyembre 14, 2026)
- Z.ai (sinuri noong Nobyembre 14, 2026)
- BroadNet Inference, pricing.json (na-sync noong Nobyembre 14, 2026) (Ingles)