Billing Units
| Model type | Common billing unit | Notes |
|---|---|---|
| Language models | USD / million tokens | Usually split into input, output, cache hit, cache write, cache read, and cache storage. |
| Multimodal models | USD / million tokens | Text, image, video, and audio can use different input/output prices. |
| Image models | USD / image or USD / million tokens | Some models expose both token-based and image-count-based pricing. |
| Video models | USD / second or USD / million tokens | Price depends on resolution, audio, video input, and editing mode. |
| Text-to-speech | USD / 10k characters | Output audio is usually not separately charged. |
| Speech-to-text | USD / second | Output text is usually not separately charged. |
| Realtime audio | USD / million tokens or USD / minute | OpenAI realtime models may price audio, text, and image inputs separately. |
/ means the source sheet did not provide a separate price, the field is not applicable, or the capability is not available.
Covered Providers
| Region | Provider | Model groups |
|---|---|---|
| Global | Volcano Engine | Seedance 2.0, seedance 2.0 fast, seedance-1.5-pro, seedream 5.0 lite, seedream 4.5, seedream 4.0, Doubao-Seed-2.0, Doubao-Seed-Character, Doubao TTS 2.0, GLM-4.7, DeepSeek-V3.2 |
| Global | Alibaba Cloud | Wan 2.5 / 2.6 / 2.7, qwen3.5 / qwen3.6, happyhorse-1.0, Qwen-Image-2.0, Qwen-3.5-Max, gui-plus-2026-02-26, qwen-flash-character-2026-02-26, z-image-turbo, qwen3-tts-flash, qwen3-asr-flash, GLM-5.1, kimi-k2.5, kimi-k2.6, DeepSeek-V3.2, DeepSeek-V4-pro, deepseek-v4-flash, MiniMax-M2.5 |
| Global | MiniMax | MiniMax-M2.7, MiniMax-Speech series, pending completion |
| Global | Kling | Kling-V3-Omni |
| Global | Vidu | Vidu Q3 series |
| Global | Xiaomi | mimo-v2.5, mimo-v2.5-pro, mimo-v2-pro |
| Global | OpenAI | GPT-5.4, GPT-Realtime series, GPT-Image-2 |
| Global | Nano banana Pro, Nano banana 2, Veo 3.1 series | |
| Global | X | Grok series, incomplete in source sheet |
| Global | Suno | suno-v3, suno-v3.5 |
| Global | ElevenLabs | Eleven v3 |
Language Models
Prices are in USD / million tokens unless otherwise stated.| Provider | Model | Context / range | Input | Output | Cache / extra pricing | RPM | Notes |
|---|---|---|---|---|---|---|---|
| MiniMax | MiniMax-M2.7 | / | 0.3000 | 1.2000 | Cache write 0.3750, cache read 0.0600 | / | |
| MiniMax | MiniMax-M2.7-highspeed | / | 0.6000 | 2.4000 | Cache write 0.3750, cache read 0.0600 | / | |
| Xiaomi | mimo-v2.5-pro | (0, 256]k | 1.0000 | 3.0000 | Cache hit 0.2000 | / | Cache write is temporarily free as of 2026-05-21 |
| Xiaomi | mimo-v2.5-pro | (256, 1024]k | 2.0000 | 6.0000 | Cache hit 0.4000 | / | |
| Xiaomi | mimo-v2-pro | (0, 256]k | 1.0000 | 3.0000 | Cache hit 0.2000 | / | |
| Xiaomi | mimo-v2-pro | (256, 1024]k | 2.0000 | 6.0000 | Cache hit 0.4000 | / | |
| Xiaomi | mimo-v2.5 | (0, 256]k | 0.4000 | 2.0000 | Cache hit 0.0800 | / | |
| Xiaomi | mimo-v2.5 | (256, 1024]k | 0.8000 | 4.0000 | Cache hit 0.1600 | / | |
| OpenAI | GPT-5.4 | Below 272k | 2.5000 | 15.0000 | Cache hit 0.2500 | / | Output price includes visible output tokens and hidden reasoning tokens. Data residency endpoints add 10%. |
| OpenAI | GPT-5.4 | Above 272k | 5.0000 | 22.5000 | Cache hit 0.5000 | / | Output price includes visible output tokens and hidden reasoning tokens. |
| X | grok-4.3 | / | 1.2500 | 2.5000 | Cache hit 0.2000 | 1,800 | 1M context window |
| X | grok-4.20-multi-agent-0309 | / | 1.2500 | 2.5000 | Cache hit 0.2000 | 1,800 | 2M context window |
| X | grok-4.20-0309-reasoning | / | 1.2500 | 2.5000 | Cache hit 0.2000 | 1,800 | 1M context window |
| X | grok-4.20-0309-non-reasoning | / | 1.2500 | 2.5000 | Cache hit 0.2000 | 1,800 | 1M context window |
| Alibaba Cloud | qwen3-max-preview | (0, 32]k | 1.2000 | 6.0000 | Cache hit 0.2400 | 600 | Input cache hits are charged at 20%. |
| Alibaba Cloud | qwen3-max-preview | (32, 128]k | 2.4000 | 12.0000 | Cache hit 0.4800 | 600 | Input cache hits are charged at 20%. |
| Alibaba Cloud | qwen3-max-preview | (128, 256]k | 3.0000 | 15.0000 | Cache hit 0.6000 | / | |
| Alibaba Cloud | qwen3.6-max-preview | (0, 128]k | 1.3000 | 7.8000 | Explicit cache hit 0.1300, explicit cache creation 1.6250 | 600 | |
| Alibaba Cloud | qwen3.6-max-preview | (128, 256]k | 2.0000 | 12.0000 | Explicit cache hit 0.2000, explicit cache creation 2.5000 | / | |
| Alibaba Cloud | qwen3.6-plus | (0, 256]k | 0.5000 | 3.0000 | Explicit cache hit 0.0500, explicit cache creation 0.6250 | 15,000 | Equivalent to qwen3.6-plus-2026-04-02 |
| Alibaba Cloud | qwen3.6-plus | (256, 1024]k | 2.0000 | 6.0000 | Explicit cache hit 0.2000, explicit cache creation 2.5000 | / | |
| Alibaba Cloud | qwen3.6-plus-2026-04-02 | (0, 256]k | 0.5000 | 3.0000 | / | 60 | Snapshot model |
| Alibaba Cloud | qwen3.6-plus-2026-04-02 | (256, 1024]k | 2.0000 | 6.0000 | / | / | Snapshot model |
| Alibaba Cloud | qwen3.5-plus | (0, 256]k | 0.4000 | 2.4000 | Explicit cache hit 0.0400, explicit cache creation 0.5000 | 15,000 | Equivalent to qwen3.5-plus-2026-02-15 |
| Alibaba Cloud | qwen3.5-plus | (256, 1024]k | 0.5000 | 3.0000 | Explicit cache hit 0.0500, explicit cache creation 0.6250 | / | |
| Alibaba Cloud | qwen3.5-plus-2026-04-20 | (0, 256]k | 0.4000 | 2.4000 | Explicit cache hit 0.0400, explicit cache creation 0.5000 | 600 | Snapshot model |
| Alibaba Cloud | qwen3.5-plus-2026-04-20 | (256, 1024]k | 0.5000 | 3.0000 | Explicit cache hit 0.0500, explicit cache creation 0.6250 | / | Snapshot model |
| Alibaba Cloud | qwen3.5-plus-2026-02-15 | (0, 256]k | 0.4000 | 2.4000 | / | 60 | Snapshot model |
| Alibaba Cloud | qwen3.5-plus-2026-02-15 | (256, 1024]k | 0.5000 | 3.0000 | / | / | Snapshot model |
| Alibaba Cloud | qwen3.6-flash | (0, 256]k | 0.2500 | 1.5000 | Explicit cache hit 0.0250, explicit cache creation 0.3125 | 15,000 | Equivalent to qwen3.6-flash-2026-04-16 |
| Alibaba Cloud | qwen3.6-flash | (256, 1024]k | 1.0000 | 4.0000 | Explicit cache hit 0.1000, explicit cache creation 1.2500 | / | |
| Alibaba Cloud | qwen3.6-flash-2026-04-16 | (0, 256]k | 0.2500 | 1.5000 | / | 60 | Snapshot model |
| Alibaba Cloud | qwen3.6-flash-2026-04-16 | (256, 1024]k | 1.0000 | 4.0000 | / | / | Snapshot model |
| Alibaba Cloud | qwen3.5-flash | / | 0.1000 | 0.4000 | Explicit cache hit 0.0100, explicit cache creation 0.1250 | 15,000 | |
| Alibaba Cloud | qwen3.5-flash-2026-02-23 | / | 0.1000 | 0.4000 | / | 60 | Snapshot model |
| Volcano Engine | doubao-seed-2.0-pro | (0, 128]k | 0.5000 | 3.0000 | Cache storage 0.0083, cache hit 0.1000 | / | |
| Volcano Engine | doubao-seed-2.0-pro | (128, 256]k | 1.0000 | 6.0000 | Cache storage 0.0083, cache hit 0.2000 | / | |
| Volcano Engine | doubao-seed-2.0-lite | (0, 128]k | 0.2500 | 2.0000 | Cache storage 0.0083, cache hit 0.0500 | / | |
| Volcano Engine | doubao-seed-2.0-lite | (128, 256]k | 0.5000 | 4.0000 | Cache storage 0.0083, cache hit 0.1000 | / | |
| Volcano Engine | doubao-seed-2.0-mini | (0, 128]k | 0.1000 | 0.4000 | Cache storage 0.0083, cache hit 0.0200 | / | |
| Volcano Engine | doubao-seed-2.0-mini | (128, 256]k | 0.2000 | 0.8000 | Cache storage 0.0083, cache hit 0.0400 | / | |
| Volcano Engine | doubao-seed-2.0-code | [0, 32]k | 0.4600 | 2.2900 | Cache storage 0.0024, cache hit 0.0914 | / | Converted from RMB pricing by dividing by 7 because overseas pricing was not available in the source. |
| Volcano Engine | doubao-seed-2.0-code | (32, 128]k | 0.6900 | 3.4300 | Cache storage 0.0024, cache hit 0.1371 | / | Converted from RMB pricing by dividing by 7. |
| Volcano Engine | doubao-seed-2.0-code | (128, 256]k | 1.3700 | 6.8600 | Cache storage 0.0024, cache hit 0.2743 | / | Converted from RMB pricing by dividing by 7. |
| Volcano Engine | Doubao-Seed-Character | [0, 32]k | 0.1100 | 0.2900 | Cache storage 0.0024, cache hit 0.0229 | / | Converted from RMB pricing by dividing by 7. |
| Volcano Engine | Doubao-Seed-Character | (32, 128]k | 0.1700 | 0.8600 | Cache storage 0.0024, cache hit 0.0229 | / | Converted from RMB pricing by dividing by 7. |
Third-party Open-weight Language Models
Prices are in USD / million tokens unless otherwise stated.| Model | Provider | Context / range | Input | Output | Cache / extra pricing | Notes |
|---|---|---|---|---|---|---|
glm-4.7 | Volcano Engine | / | 0.6000 | 2.2000 | Cache storage 0.0083, explicit cache creation 0.1100 | |
glm-5.1 | Bailian | (0, 32]k | 0.825 | 3.301 | / | Non-thinking and thinking modes. Output includes chain-of-thought and answer. Free quota: 1M tokens for 90 days. |
glm-5.1 | Bailian | (32, 200]k | 1.100 | 3.851 | / | Non-thinking and thinking modes. |
kimi-k2.5 | Bailian | / | 0.574 | 3.011 | / | |
kimi-k2.6 | Bailian | / | 0.8939 | 3.7131 | / | |
DeepSeek-V3.2 | Volcano Engine | (0, 32]k | 0.2800 | 0.4200 | Cache storage 0.0083, cache creation 0.0560 | |
DeepSeek-V3.2 | Volcano Engine | (32, 128]k | 0.5600 | 0.8400 | Cache storage 0.0083, cache creation 0.0560 | |
DeepSeek-V3.2 | Bailian | / | 0.5700 | 1.7100 | Explicit cache creation 0.7130, implicit cache hit 0.1140, explicit cache hit 0.0570 | Non-thinking and thinking modes. Output includes chain-of-thought and answer. |
DeepSeek-V4-pro | Bailian | / | 2.40 | 4.80 | Cache hit 0.20 | |
deepseek-v4-flash | Bailian | / | 0.20 | 0.40 | Cache hit 0.04 | |
MiniMax-M2.5 | Bailian | / | 0.3040 | 1.2130 | / | Thinking mode only. Output includes chain-of-thought and answer. |
Open-weight Qwen Series
| Provider | Model | Context / range | Input | Output |
|---|---|---|---|---|
| Alibaba Cloud | qwen3.5-397b-a17b | (0, 256]k | 0.600 | 3.600 |
| Alibaba Cloud | qwen3.5-27b | (0, 256]k | 0.300 | 2.400 |
| Alibaba Cloud | qwen3.6-35b-a3b | (0, 256]k | 0.248 | 1.485 |
| Alibaba Cloud | qwen3.6-27b | (0, 256]k | 0.600 | 3.600 |
| Alibaba Cloud | qwen3.5-122b-a10b | (0, 256]k | 0.400 | 3.200 |
| Alibaba Cloud | qwen3.5-35b-a3b | (0, 256]k | 0.250 | 2.000 |
Video Models
Video prices are listed in USD / second unless the unit column says USD / million tokens.| Provider | Model | Audio | Resolution / mode | Input contains video | Price | Unit | Input price | Capability / notes |
|---|---|---|---|---|---|---|---|---|
Veo 3.1 | Audio | 720p / 1080p | No | 0.400 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 | Audio | 4K | No | 0.600 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 | No audio | 720p / 1080p | No | 0.200 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 | No audio | 4K | No | 0.400 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | Audio | 720p | No | 0.100 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | Audio | 1080p | No | 0.120 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | Audio | 4K | No | 0.300 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | No audio | 720p | No | 0.080 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | No audio | 1080p | No | 0.100 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Fast | No audio | 4K | No | 0.250 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Lite | Audio | 720p | No | 0.050 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Lite | Audio | 1080p | No | 0.080 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Lite | No audio | 720p | No | 0.030 | Second | / | Image-to-video, text-to-video | |
Veo 3.1 Lite | No audio | 1080p | No | 0.050 | Second | / | Image-to-video, text-to-video | |
| X | grok-imagine-video | / | 480p | Yes, charged separately | 0.050 | Second | 0.01 USD / second and 0.002 USD / image | Image-to-video, text-to-video, video-to-video; 70 RPM |
| X | grok-imagine-video | / | 720p | Yes, charged separately | 0.070 | Second | Same as above | Image-to-video, text-to-video, video-to-video; 70 RPM |
| Volcano Engine | Seedance 2.0 | / | 480p / 720p | No | 7.000 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | Seedance 2.0 | / | 480p / 720p | Yes | 4.300 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | Seedance 2.0 | / | 1080p | No | 7.700 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | Seedance 2.0 | / | 1080p | Yes | 4.700 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | seedance 2.0 fast | / | 480p / 720p | No | 5.600 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | seedance 2.0 fast | / | 480p / 720p | Yes | 3.300 | Million tokens | / | Image-to-video, text-to-video, video editing |
| Volcano Engine | seedance-1.5-pro | Audio | 480p / 720p / 1080p | No | 2.400 | Million tokens | / | Image-to-video, text-to-video; 600 RPM |
| Volcano Engine | seedance-1.5-pro | No audio | 480p / 720p / 1080p | No | 1.200 | Million tokens | / | Image-to-video, text-to-video; 600 RPM |
| Alibaba Cloud | wan2.6-i2v-flash | Audio | 720p / 1080p | No | 0.0500 / 0.0750 | Second | / | Image-to-video based on first frame; 300 RPM |
| Alibaba Cloud | wan2.6-i2v-flash | No audio | 720p / 1080p | No | 0.0250 / 0.0375 | Second | / | Image-to-video based on first frame; 300 RPM |
| Alibaba Cloud | wan2.6-r2v-flash | Audio | 720p / 1080p | Yes | 0.0500 / 0.0750 | Second | / | Reference-to-video; 300 RPM |
| Alibaba Cloud | wan2.6-r2v-flash | No audio | 720p / 1080p | Yes | 0.0250 / 0.0375 | Second | / | Reference-to-video; 300 RPM |
| Alibaba Cloud | wan2.6-i2v | Audio | 720p / 1080p | No | 0.1000 / 0.1500 | Second | / | Image-to-video based on first frame; 300 RPM |
| Alibaba Cloud | wan2.6-r2v | No audio | 720p / 1080p | Yes | 0.1000 / 0.1500 | Second | / | Reference-to-video; 300 RPM |
| Alibaba Cloud | wan2.6-t2v | No audio | 720p / 1080p | No | 0.1000 / 0.1500 | Second | / | Text-to-video; 300 RPM |
| Alibaba Cloud | wan2.5-i2v-preview | No audio | 480p / 720p / 1080p | No | 0.0500 / 0.1000 / 0.1500 | Second | / | Image-to-video; 300 RPM |
| Alibaba Cloud | wan2.5-t2v-preview | No audio | 480p / 720p / 1080p | No | 0.0500 / 0.1000 / 0.1500 | Second | / | Text-to-video; 300 RPM |
| Alibaba Cloud | wan2.7-r2v | No audio | 720p / 1080p | Yes | 0.1000 / 0.1500 | Second | / | Reference-to-video, supports image, text, audio, and video input; 300 RPM |
| Alibaba Cloud | wan2.7-i2v-2026-04-25, wan2.7-i2v | No audio | 720p / 1080p | No | 0.1000 / 0.1500 | Second | / | Image-to-video, supports image, text, and audio input; 300 RPM |
| Alibaba Cloud | wan2.7-t2v-2026-04-25, wan2.7-t2v | No audio | 720p / 1080p | No | 0.1000 / 0.1500 | Second | / | Text-to-video, supports text and audio input; 300 RPM |
| Alibaba Cloud | wan2.7-videoedit | No audio | 720p / 1080p | Yes | 0.1000 / 0.1500 | Second | / | Video editing, supports image, text, and video input; 300 RPM |
| Alibaba Cloud | happyhorse-1.0-t2v | No audio | 720p / 1080p | No | 0.1400 / 0.2400 | Second | / | Text-to-video only; 20% limited-time discount on 2026.5.21; 300 RPM |
| Alibaba Cloud | happyhorse-1.0-i2v | No audio | 720p / 1080p | No | 0.1400 / 0.2400 | Second | / | Image-to-video; 20% limited-time discount on 2026.5.21; 300 RPM |
| Alibaba Cloud | happyhorse-1.0-r2v | No audio | 720p / 1080p | No | 0.1400 / 0.2400 | Second | / | Reference-to-video; 20% limited-time discount on 2026.5.21; 300 RPM |
| Alibaba Cloud | happyhorse-1.0-video-edit | No audio | 720p / 1080p | Yes | 0.1400 / 0.2400 | Second | / | Video editing; 20% limited-time discount on 2026.5.21; 300 RPM |
| Kling | Kling-V3-Omni | No audio | std / pro / 4K | No | 0.0840 / 0.1120 / 0.4200 | Second | / | Video generation |
| Kling | Kling-V3-Omni | Audio | std / pro / 4K | No | 0.1120 / 0.1400 / 0.4200 | Second | / | Video generation |
| Kling | Kling-V3-Omni | No audio | std / pro | Yes | 0.1260 / 0.1680 | Second | / | Video editing |
| Vidu | Vidu Q3-pro | / | 1080p / 720p / 540p | No | 0.1200 / 0.1000 / 0.0450 | Second | / | Image-to-video, text-to-video, first/last frame; 1-16 seconds |
| Vidu | Vidu Q3-turbo | / | 1080p / 720p / 540p | No | 0.6500 / 0.0550 / 0.0350 | Second | / | Image-to-video, text-to-video, first/last frame; 1-16 seconds |
| Vidu | Vidu Q3-pro-fast | / | 1080p / 720p | No | 0.1250 / 0.1000 | Second | / | Image-to-video; 1-16 seconds |
| Vidu | Vidu Q3-mix | / | 720p / 1080p | No | 0.1200 / 0.1450 | Second | / | Reference-to-video; 3-16 seconds |
| Vidu | Vidu Q3-turbo | / | 540p / 720p / 1080p | No | 0.0200 / 0.0500 / 0.0650 | Second | / | Reference-to-video; 3-16 seconds |
| Vidu | Vidu Q3 | / | 540p / 720p / 1080p | No | 0.0350 / 0.0600 / 0.0750 | Second | / | Reference-to-video; 3-16 seconds |
Image Models
Image models can be priced by tokens and/or images depending on the provider.| Provider | Model | Input type | Output mode | Input USD / million tokens | Output USD / million tokens | Cached input | Input USD / image | Output USD / image | RPM / notes |
|---|---|---|---|---|---|---|---|---|---|
| OpenAI | GPT-Image-2 | Text | / | 5.000 | / | 1.250 | / | / | |
| OpenAI | GPT-Image-2 | Image | / | 8.000 | 30.000 | 2.000 | / | / | |
Nano banana Pro | Text | / | 2.000 | 12.000 | / | / | / | Shared free quota: 5,000 search-enhanced calls per month across Gemini 3 models; then 14 USD / 1,000 calls. | |
Nano banana Pro | Image | 1K / 2K | 2.000 | 120.000 | / | 0.001 | 0.134 | ||
Nano banana Pro | Image | 4K | 2.000 | 120.000 | / | 0.001 | 0.240 | ||
Nano banana 2 | Text | / | 0.500 | 3.000 | / | / | / | ||
Nano banana 2 | Image | 0.5K | 0.500 | 60.000 | / | / | 0.045 | ||
Nano banana 2 | Image | 1K | 0.500 | 60.000 | / | / | 0.067 | ||
Nano banana 2 | Image | 2K | 0.500 | 60.000 | / | / | 0.101 | ||
Nano banana 2 | Image | 4K | 0.500 | 60.000 | / | / | 0.151 | ||
| X | grok-imagine-image | Text / image | 1K | / | / | / | 0.002 | 0.020 | 300 RPM |
| X | grok-imagine-image | Text / image | 2K | / | / | / | / | 0.020 | |
| X | grok-imagine-image-quality | Text / image | 1K | / | / | / | 0.010 | 0.050 | 300 RPM |
| X | grok-imagine-image-quality | Text / image | 2K | / | / | / | / | 0.070 | |
| Alibaba Cloud | wan2.6-image | Text / image | / | / | / | / | / | 0.030 | 300 RPM |
| Alibaba Cloud | wan2.6-t2i | Text | / | / | / | / | / | 0.030 | 300 RPM |
| Alibaba Cloud | wan2.5-t2i-preview | Text | / | / | / | / | / | 0.030 | 300 RPM |
| Alibaba Cloud | z-image-turbo | Text | Standard | / | / | / | / | 0.150 | 120 RPM |
| Alibaba Cloud | z-image-turbo | Text | Thinking | / | / | / | / | 0.030 | |
| Alibaba Cloud | wan2.7-image-pro | Text / image | / | / | / | / | / | 0.075 | 300 RPM |
| Alibaba Cloud | wan2.7-image | Text / image | / | / | / | / | / | 0.030 | 300 RPM |
| Alibaba Cloud | qwen-image-2.0-pro-2026-04-22 | Text / image | / | / | / | / | / | 0.075 | 2 RPM |
| Alibaba Cloud | qwen-image-2.0 | Text / image | / | / | / | / | / | 0.035 | 120 RPM |
| Volcano Engine | seedream 5.0 lite | Text / image | / | / | / | / | / | 0.035 | |
| Volcano Engine | seedream 4.5 | Text / image | / | / | / | / | / | 0.040 | |
| Volcano Engine | seedream 4.0 | Text / image | / | / | / | / | / | 0.030 |
Multimodal Models
Prices are in USD / million tokens.| Provider | Model | Input: text / image / video | Input: audio | Output: text | Output: text + audio | Series | Notes |
|---|---|---|---|---|---|---|---|
| Alibaba Cloud | qwen3.5-omni-plus | 1.40 | 11.00 | 8.30 | 44.00 | Qwen Omni | Each 1M token free quota is valid for 90 days; equivalent to qwen3.5-omni-plus-2026-03-15 |
| Alibaba Cloud | qwen3.5-omni-plus-2026-03-15 | 1.40 | 11.00 | 8.30 | 44.00 | Qwen Omni | Snapshot model |
| Alibaba Cloud | qwen3.5-omni-flash | 0.40 | 3.00 | 2.20 | 11.90 | Qwen Omni | Equivalent to qwen3.5-omni-flash-2026-03-15 |
| Alibaba Cloud | qwen3.5-omni-flash-2026-03-15 | 0.40 | 3.00 | 2.20 | 11.90 | Qwen Omni | Snapshot model |
| Alibaba Cloud | qwen3.5-omni-plus-realtime | 2.10 | 16.50 | 12.40 | 62.00 | Qwen Omni-Realtime | Each 1M token free quota is valid for 90 days |
| Alibaba Cloud | qwen3.5-omni-plus-realtime-2026-03-15 | 2.10 | 16.50 | 12.40 | 62.00 | Qwen Omni-Realtime | Snapshot model |
| Alibaba Cloud | qwen3.5-omni-flash-realtime | 0.55 | 4.50 | 3.30 | 17.70 | Qwen Omni-Realtime | |
| Alibaba Cloud | qwen3.5-omni-flash-realtime-2026-03-15 | 0.55 | 4.50 | 3.30 | 17.70 | Qwen Omni-Realtime | Snapshot model |
Realtime and Audio Models
Realtime conversation
| Provider | Model | Input type | Input price | Output price | Cached input | Output type |
|---|---|---|---|---|---|---|
| OpenAI | GPT-Realtime-2 | Audio | 32.00 USD / million tokens | 64.00 USD / million tokens | 0.40 | Text, audio |
| OpenAI | GPT-Realtime-2 | Text | 4.00 USD / million tokens | 24.00 USD / million tokens | 0.40 | Text, audio |
| OpenAI | GPT-Realtime-2 | Image | 5.00 USD / million tokens | / | 0.50 | Text, audio |
Text-to-speech
| Provider | Model | Price | Capability | Notes |
|---|---|---|---|---|
| ElevenLabs | Eleven v3 | 1.00 USD / 10k characters | Text-to-speech | |
| Alibaba Cloud | qwen3-tts-flash-realtime | 0.13 USD / 10k characters | Text-to-speech | Equivalent to qwen3-tts-flash-realtime-2025-11-27 |
| Alibaba Cloud | qwen3-tts-flash-realtime-2025-11-27 | 0.13 USD / 10k characters | Text-to-speech | Snapshot model |
| Alibaba Cloud | qwen3-tts-flash-realtime-2025-09-18 | 0.13 USD / 10k characters | Text-to-speech | Snapshot model |
| Alibaba Cloud | qwen3-tts-flash | 0.10 USD / 10k characters | Text-to-speech | Equivalent to qwen3-tts-flash-2025-11-27 |
| Alibaba Cloud | qwen3-tts-flash-2025-11-27 | 0.10 USD / 10k characters | Text-to-speech | Snapshot model |
| Alibaba Cloud | qwen3-tts-flash-2025-09-18 | 0.10 USD / 10k characters | Text-to-speech | Snapshot model |
| Volcano Engine | Doubao-Text-to-Speech-2.0 | 0.4286 USD / 10k characters | Text-to-speech | Converted from RMB pricing because overseas pricing was unavailable due to account-region restrictions. |
Speech-to-text
| Provider | Model | Price | Capability | Notes |
|---|---|---|---|---|
| Alibaba Cloud | qwen3-asr-flash | 0.000035 USD / second | Speech-to-text | Equivalent to qwen3-asr-flash-2025-09-08 |
| Alibaba Cloud | qwen3-asr-flash-2026-02-10 | 0.000035 USD / second | Speech-to-text | Snapshot model |
| Alibaba Cloud | qwen3-asr-flash-2025-09-08 | 0.000035 USD / second | Speech-to-text | Snapshot model |
| Alibaba Cloud | qwen3-asr-flash-filetrans | 0.000035 USD / second | Speech-to-text | Main model |
| Alibaba Cloud | qwen3-asr-flash-filetrans-2025-11-17 | 0.000035 USD / second | Speech-to-text | Snapshot model |
| Alibaba Cloud | qwen3-asr-flash-realtime | 0.000090 USD / second | Speech-to-text | Equivalent to qwen3-asr-flash-realtime-2025-10-27 |
| Alibaba Cloud | qwen3-asr-flash-realtime-2026-02-10 | 0.000090 USD / second | Speech-to-text | Snapshot model |
| Alibaba Cloud | qwen3-asr-flash-realtime-2025-10-27 | 0.000090 USD / second | Speech-to-text | Snapshot model |
Realtime translation and streaming transcription
| Provider | Model | Input type | Output type | Price | Capability |
|---|---|---|---|---|---|
| OpenAI | GPT-Realtime-Translate | Audio | Audio, text | 0.034 USD / minute or 0.00057 USD / second | Realtime translation |
| OpenAI | GPT-Realtime-Whisper | Audio, text | Text | 0.017 USD / minute or 0.00028 USD / second | Realtime streaming speech-to-text |
Other Models
| Type | Provider | Model | Input price | Output price | Notes |
|---|---|---|---|---|---|
| GUI interaction | Alibaba Cloud | gui-plus-2026-02-26 | 0.21 USD / million tokens | 0.64 USD / million tokens | Converted from RMB pricing by dividing by 7 because overseas pricing was not available. |
| Character roleplay | Alibaba Cloud | qwen-flash-character-2026-02-26 | 0.05 USD / million tokens | 0.40 USD / million tokens | Each 1M token free quota is valid for 90 days. |
| Music generation | Suno | suno-v3 | 0.30 USD / call | / | The source notes Suno does not provide an official API platform; price is referenced from api4gpt.com. |
| Music generation | Suno | suno-v3.5 | 0.30 USD / call | / | The source notes Suno does not provide an official API platform; price is referenced from api4gpt.com. |
Pricing Tips
- For language models, confirm context-length tiers and cache behavior before estimating cost.
- For video models, check resolution, duration, audio mode, whether video input is used, and whether the request is generation or editing.
- For image models, verify whether the provider charges by image, by tokens, or both.
- For realtime and audio models, check whether the unit is per million tokens, per second, per minute, or per 10k characters.
- Free quotas, discounts, and converted prices are provider-specific and should be treated as estimates unless confirmed in the current provider console.