ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com

Global & China LLM API Official Rates & Task Costs

Drop real CSV/JSONL usage logs to replay bills, customize monthly token volumes, and pin up to 4 endpoints to export Markdown.

Buy GPU vs Call API Break-Even →

Sorted by weighted blended /1M rate (70% input + 30% output)· Verified 2026-11-20

Model & ProviderRelative TierInput / Cached (/1M)Output (/1M)Access & SpecsPin / Link
Qwen3.5-4B (免费调用)
SiliconFlow·262K
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON·Vision

100% Free tier (rate-limited)

Open
Qwen3-8B (免费调用)
SiliconFlow·131K
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON

100% Free tier (rate-limited)

Open
GLM-4.7-Flash (免费旗舰轻量)
Zhipu BigModel·200K
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON

Official 100% free tier from Zhipu AI

Open
GLM-4.6V-Flash (免费多模态)
Zhipu BigModel·128K
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON·Vision

Official 100% free tier from Zhipu AI

Open
ERNIE-5.5-Flash (1M 免费预览)
Baidu Qianfan·1M
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON·Vision

Free preview tier for Baidu ERNIE 5.5 Flash (1M ctx)

Open
Gemma 4 31B IT (免费路由)
OpenRouter·262K
FREE
FREE $0
FREE $0
CN Direct·Tools·Vision

Free rate-limited developer endpoint on OpenRouter

Open
Ant Ling 3.2 Flash (免费路由)
OpenRouter·1M
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON·Vision

Free rate-limited developer endpoint on OpenRouter

Open
Ling 3.2 Flash
Ant 蚂蚁百灵·1M
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON·Vision

Free dev/experiment tier · API ID: ling-3.2-flash

Open
North Small Translate
Cohere·33K
FREE
FREE $0
FREE $0
Global

Free dev/experiment tier · API ID: north-small-translate-09-2026

Open
Leanstral 1.5
Mistral AI·262K
FREE
FREE $0
FREE $0
Global·Tools·JSON·Vision

Free dev/experiment tier · API ID: labs-leanstral-1-5-1

Open
North Mini Code
Cohere·256K
FREE
FREE $0
FREE $0
Global·Tools

Free dev/experiment tier · API ID: north-mini-code-1-0

Open
Nemotron 3 Nano Omni
NVIDIA·256K
FREE
FREE $0
FREE $0
Global·Tools·Vision

Free dev/experiment tier · API ID: nemotron-3-nano-omni-30b-a3b-reasoning

Open
Devstral Small 2
Mistral AI·256K
FREE
FREE $0
FREE $0
Global·Tools·JSON·Vision

Free dev/experiment tier · API ID: labs-devstral-small-2512

Open
GLM-4.5-Flash
Zhipu BigModel·131K
FREE
FREE $0
FREE $0
CN Direct·Tools·JSON

Free dev/experiment tier · API ID: glm-4.5-flash

Open
Ling 3.0 Tiny
Ant 蚂蚁百灵·262K
$0.024/M
Lowest Paid
$0.015
Cache: $0.0030
$0.045
CN Direct·Tools·JSON

API ID: ling-3.0-tiny

Open
Llama-3.1-8B-Instruct
Meta AI·131K
$0.026/M
1.1x
$0.020
Cache: $0.0040
$0.040
Global·Tools·JSON

API ID: llama-3.1-8b-instruct

Open
Ling 3.0 Flash
Ant 蚂蚁百灵·262K
$0.034/M
1.4x
$0.021
Cache: $0.0042
$0.063
CN Direct·Tools

API ID: ling-3.0-flash

Open
Doubao-Flash (豆包极速轻量版)
Volcengine Ark·128K
$0.040/M
1.6x
$0.021
Cache: $0.0042
$0.083
CN Direct·Tools·JSON

Starts at ¥0.15/1M input on Volcengine Ark

Open
Qwen3.7-Flash
Alibaba Bailian·1M
$0.053/M
2.2x
$0.028
Cache: $0.0056
$0.11
CN Direct·Tools·JSON·Vision

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Qwen-Turbo (极速轻量)
Alibaba Bailian·1M
$0.056/M
2.3x
$0.043
Cache: $0.0083
$0.085
CN Direct·Tools·JSON

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Hunyuan Hy4-Flash (混元 4 极速)
Tencent Hunyuan·1M
$0.066/M
2.7x
$0.035
Cache: $0.0014
$0.14
CN Direct·Tools·JSON·Vision

Tencent Cloud Hunyuan official API

Open
Amazon Nova Micro 1.0
Amazon Bedrock·128K
$0.067/M
2.8x
$0.035
Cache: $0.0088
$0.14
Global·Tools

AWS Bedrock official on-demand token pricing

Open
Nova Micro 1.0
Amazon Bedrock·128K
$0.067/M
2.8x
$0.035
Cache: $0.0070
$0.14
Global·Tools

API ID: nova-micro-v1

Open
Command R7B Arabic
Cohere·128K
$0.071/M
3.0x
$0.037
Cache: $0.0074
$0.15
Global·Tools·JSON

API ID: command-r7b-arabic-02-2025

Open
Command R7B
Cohere·128K
$0.071/M
3.0x
$0.037
Cache: $0.037
$0.15
Global·Tools·JSON

Enterprise RAG & multilingual API with free Trial tier

Open
Ant Ling 3.1 Flash
OpenRouter·1M
$0.072/M
3.0x
$0.045
Cache: $0.0090
$0.14
CN Direct·Tools

Unified multi-provider routing via OpenRouter

Open
Qwen-Flash (超低价走量)
Alibaba Bailian·1M
$0.077/M
3.2x
$0.021
Cache: $0.0042
$0.21
CN Direct·Tools·JSON

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Seed 1.6 Flash
Volcengine Ark·256K
$0.077/M
3.2x
$0.021
Cache: $0.0042
$0.21
CN Direct·Tools·JSON·Vision

API ID: doubao-seed-1-6-flash-250828

Open
Hy-MT2-1.8B
Tencent Hunyuan·8K
$0.084/M
3.5x
$0.044
Cache: $0.0088
$0.18
CN Direct

API ID: hy-mt2-1.8b

Open
DeepSeek-R1-Distill-Qwen-7B
Alibaba Bailian·33K
$0.090/M
3.8x
$0.069
Cache: $0.014
$0.14
CN Direct·Tools·JSON

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Microsoft Phi-4 (14B)
OpenRouter·16K
$0.091/M
3.8x
$0.070
Cache: $0.070
$0.14
CN Direct·JSON

Unified multi-provider routing via OpenRouter

Open
Phi 4
Microsoft·16K
$0.091/M
3.8x
$0.070
Cache: $0.014
$0.14
Global·JSON

API ID: phi-4

Open
Ling 3.0 Flash VL
Ant 蚂蚁百灵·131K
$0.096/M
4.0x
$0.060
Cache: $0.012
$0.18
CN Direct·Tools·JSON·Vision

API ID: ling-3.0-flash-vl

Open
Ministral 3 3B
Mistral AI·131K
$0.100/M
4.2x
$0.10
Cache: $0.010
$0.10
Global·Tools·JSON·Vision

Mistral La Plateforme API (includes free experiment tier)

Open
Doubao-Seed-2.0-Mini
Volcengine Ark·256K
$0.10/M
4.3x
$0.028
Cache: $0.0056
$0.28
CN Direct·Tools·JSON·Vision

Volcengine Ark high-concurrency CN endpoint

Open
Nemotron 3.5 Lightning 30B A3B
NVIDIA·1M
$0.11/M
4.5x
$0.070
Cache: $0.040
$0.20
Global·Tools·JSON

API ID: nemotron-3.5-lightning

Open
Amazon Nova Lite 1.0
Amazon Bedrock·300K
$0.11/M
4.7x
$0.060
Cache: $0.015
$0.24
Global·Tools·Vision

AWS Bedrock official on-demand token pricing

Open
Nemotron 3 Nano 30B A3B
NVIDIA·262K
$0.11/M
4.7x
$0.060
Cache: $0.012
$0.24
Global·Tools·JSON

API ID: nemotron-3-nano-30b-a3b

Open
Nova Lite 1.0
Amazon Bedrock·300K
$0.11/M
4.7x
$0.060
Cache: $0.012
$0.24
Global·Tools·Vision

API ID: nova-lite-v1

Open
LFM3-1.7B
Liquid AI·262K
$0.12/M
4.8x
$0.080
Cache: $0.020
$0.20
Global·Tools·JSON

API ID: lfm-3-1.7b

Open
Qwen-Omni Turbo
Alibaba Bailian·33K
$0.13/M
5.4x
$0.070
Cache: $0.014
$0.27
CN Direct·Tools·JSON·Vision

API ID: qwen-omni-turbo

Open
Qwen-Long (千万长文档)
Alibaba Bailian·1M
$0.13/M
5.5x
$0.069
Cache: $0.014
$0.28
CN Direct·JSON

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Qwen3-14B (硅基托管)
SiliconFlow·131K
$0.13/M
5.5x
$0.069
Cache: $0.069
$0.28
CN Direct·Tools·JSON

Open-weight cloud hosting on SiliconFlow

Open
Llama 3.2 3B Instruct
Meta AI·131K
$0.13/M
5.6x
$0.050
Cache: $0.010
$0.33
Global·JSON

API ID: llama-3.2-3b-instruct

Open
Hy-MT2-7B
Tencent Hunyuan·8K
$0.14/M
5.8x
$0.074
Cache: $0.015
$0.29
CN Direct·JSON

API ID: hy-mt2-7b

Open
Hy-MT2-30B-A3B
Tencent Hunyuan·8K
$0.14/M
5.8x
$0.074
Cache: $0.015
$0.29
CN Direct·JSON

API ID: hy-mt2-30b-a3b

Open
Ministral 3 8B
Mistral AI·262K
$0.15/M
6.3x
$0.15
Cache: $0.015
$0.15
Global·Tools·JSON·Vision

Mistral La Plateforme API (includes free experiment tier)

Open
Pixtral 12B
Mistral AI·128K
$0.15/M
6.3x
$0.15
Cache: $0.030
$0.15
Global·Tools·JSON·Vision

API ID: pixtral-12b

Open
Open Mistral Nemo
Mistral AI·128K
$0.15/M
6.3x
$0.15
Cache: $0.030
$0.15
Global·Tools·JSON

API ID: open-mistral-nemo

Open
solar-mini
Upstage·33K
$0.15/M
6.3x
$0.15
Cache: $0.030
$0.15
Global·Tools·JSON

API ID: solar-mini

Open
Gemma 4 26B-A4B IT
OpenRouter·262K
$0.15/M
6.4x
$0.090
Cache: $0.050
$0.30
CN Direct·Tools·JSON·Vision

Unified multi-provider routing via OpenRouter

Open
GPT-5 Nano
OpenAI Official·400K
$0.15/M
6.5x
$0.050
Cache: $0.0050
$0.40
Global·Tools·JSON·Vision

API ID: gpt-5-nano

Open
Step-3.5-Flash
StepFun·256K
$0.16/M
6.5x
$0.097
Cache: $0.019
$0.29
CN Direct·Tools·JSON

Official StepFun CN platform

Open
Llama 4 Scout (109B MoE)
OpenRouter·328K
$0.16/M
6.7x
$0.10
Cache: $0.10
$0.30
CN Direct·Tools·JSON·Vision

Unified multi-provider routing via OpenRouter

Open
Mistral Small 3.2
Mistral AI·128K
$0.16/M
6.7x
$0.10
Cache: $0.020
$0.30
Global·Tools·JSON·Vision

API ID: mistral-small-2506

Open
Llama-4-Scout-109B-A17B (MoE)
Meta AI·1M
$0.16/M
6.7x
$0.10
Cache: $0.10
$0.30
CN Direct·Tools·JSON·Vision

OpenRouter: Unified multi-provider routing via OpenRouter

Open
Qwen-Plus (高性价比主力)
Alibaba Bailian·1M
$0.16/M
6.7x
$0.11
Cache: $0.022
$0.28
CN Direct·Tools·JSON

Hosted on Alibaba Cloud Bailian (CN Direct)

Open
Seed Character
Volcengine Ark·256K
$0.16/M
6.7x
$0.11
Cache: $0.022
$0.28
CN Direct·Tools·JSON·Vision

API ID: doubao-seed-character-260628

Open
Qwen3.5-35B-A3B (硅基托管)
SiliconFlow·262K
$0.17/M
7.2x
$0.056
Cache: $0.056
$0.44
CN Direct·Tools·JSON·Vision

Open-weight cloud hosting on SiliconFlow

Open
Nemotron 3 Super 120B A12B
NVIDIA·262K
$0.17/M
7.2x
$0.075
Cache: $0.015
$0.40
Global·Tools·JSON

API ID: nemotron-3-super-120b-a12b

Open

Intelligence vs. Cost / VRAM Pareto Quadrant

Pareto Frontier

Top-left corner = maximum benchmark capability for minimum VRAM or API spend

↖ Sweet-Spot Value Zone¥0.25/M¥1/M¥3/M¥10/M¥35/M¥120/M1528415366Qwen3.8-FlashQwen3.5-397B-A17BDeepSeek-V4.1-FlashStep-3.5-FlashGemini 4 NeonGemini 4 HydrogenLing 3.0 Flash
Qwen3.5-397B-A17BAlibaba Bailian综合 ¥3/MIndex 51Pareto Optimal
Pareto China GlobalInspect →