AIUpdateWatch Daily Intelligence
A practical global comparison. Main Chinese models are part of the core tracker, with facts, provider claims and AIUpdateWatch interpretation kept separate.
Prices remain in their source currency. A CNY price is never relabeled as USD, and region-specific access is kept visible.
| Origin | Provider | Model | Released | Availability | Context | Input | Output |
|---|---|---|---|---|---|---|---|
| International | OpenAI | GPT-5.6 Sol | 2026-07-09 | API and eligible ChatGPT plans | See current model documentation (opens in a new tab) | $5/1M | $30/1M |
| International | OpenAI | GPT-5.6 Terra | 2026-07-09 | API and eligible ChatGPT plans | See current model documentation (opens in a new tab) | $2.5/1M | $15/1M |
| International | OpenAI | GPT-5.6 Luna | 2026-07-09 | API and eligible ChatGPT plans | See current model documentation (opens in a new tab) | $1/1M | $6/1M |
| International | Anthropic | Claude Fable 5 | 2026-06 | Claude and API where available | See current model documentation (opens in a new tab) | $10/1M | $50/1M |
| International | xAI | Grok 4.5 | 2026-07-16 | xAI API and Grok products | 500K tokens (opens in a new tab) | $2/1M | $6/1M |
| International | Gemini 3.5 Flash | 2026-05-19 | Gemini ecosystem and developer services | See current model documentation (opens in a new tab) | Verify | Verify | |
| China | DeepSeek | V4 Flash | 2026-04-24 preview | DeepSeek API via deepseek-v4-flash or deepseek-v4-pro; legacy aliases retire July 24 at 15:59 UTC | 1M tokens (opens in a new tab) | Verify | Verify |
| China | Alibaba Cloud | Qwen3.7-Max | 2026 | Alibaba Cloud Model Studio; regional endpoints vary | See current Model Studio catalog (opens in a new tab) | Verify | Verify |
| China | Moonshot AI | Kimi K3 | 2026-07 | Kimi API platform; region and account eligibility vary | 1M tokens (opens in a new tab) | ¥20/1M | ¥100/1M |
| China | Zhipu AI | GLM-5.2 | 2026 | Zhipu BigModel platform; regional access varies | 1M tokens (opens in a new tab) | Verify | Verify |
| China | Baidu | ERNIE 5.0 | 2026-01 | Baidu Qianfan and ERNIE services; international catalog availability varies | 128K tokens (opens in a new tab) | $1.4/1M | $5.6/1M |
| China | ByteDance | Doubao Seed 2.1 | 2026 | Volcengine Ark; regional availability varies | See current Volcengine Ark catalog (opens in a new tab) | Verify | Verify |
| China | MiniMax | MiniMax-M3 | 2026-06-01 | MiniMax API platform and compatible endpoints | 1M tokens (opens in a new tab) | Verify | Verify |
| China | StepFun | Step 3.7 Flash | 2026 | StepFun open platform; regional access varies | See current StepFun model catalog (opens in a new tab) | ¥1.35/1M | ¥8.1/1M |
| China | Tencent | Hunyuan A13B | 2025-06-25 | Tencent Cloud Hunyuan / migration path to TokenHub | 224K max input; 32K max output (opens in a new tab) | Verify | Verify |
| China | Alibaba Cloud | Qwen-Image-3.0 | 2026-07-22 | Qwen and Alibaba Cloud ecosystem; verify regional endpoint availability | Up to 4.5K-token image instructions (opens in a new tab) | Verify | Verify |
| China | Alibaba Cloud | Qwen-Audio-3.0-TTS Flash / Plus | 2026-07-21 | Alibaba Cloud Model Studio; rollout and regional availability may vary | 16-language text-to-speech release (opens in a new tab) | Verify | Verify |
International providers
Best fit: Premium reasoning and agentic work
Modalities: Text, vision and tool-enabled workflows
Context: See current model documentation
Important caveat: Launch comparisons are provider-published; confirm task-specific performance independently.
Open official source (opens in a new tab)Best fit: Mid-tier capability and cost balance
Modalities: Text, vision and tool-enabled workflows
Context: See current model documentation
Important caveat: Use current API documentation for limits and regional availability.
Open official source (opens in a new tab)Best fit: Lowest-cost member of the GPT-5.6 family
Modalities: Text and general assistant workflows
Context: See current model documentation
Important caveat: Lower price does not imply best performance for every workload.
Open official source (opens in a new tab)Best fit: Premium long-running agent work
Modalities: Text, vision, coding and long-running agents
Context: See current model documentation
Important caveat: Anthropic documents additional safeguards and retention behavior for Mythos-class traffic.
Open official source (opens in a new tab)Best fit: Coding, agentic tasks and cost efficiency
Modalities: Text, code, web/X search and tools
Context: 500K tokens
Important caveat: Requests exceeding 200K context use higher prices: $4 input and $12 output per 1M tokens.
Open official source (opens in a new tab)Best fit: Fast agentic and multimodal workflows
Modalities: Multimodal, coding and interactive UI generation
Context: See current model documentation
Important caveat: This edition did not capture a stable region-neutral official API price; verify before purchase.
Open official source (opens in a new tab)China model watch
DeepSeek, Alibaba Qwen, Moonshot Kimi, Zhipu GLM, Baidu ERNIE, ByteDance Doubao, MiniMax, StepFun and Tencent Hunyuan are permanent first-class entries in the model, benchmark, pricing and source-monitoring systems.
Best fit: Large-context efficiency and open-model ecosystem
Modalities: Text, reasoning and coding
Context: 1M tokens
Published input: Verify
Published output: Verify
Caveat: Production integrations should use the explicit V4 model names. deepseek-chat and deepseek-reasoner are scheduled to stop working after the retirement deadline.
Open official source (opens in a new tab)Best fit: Complex multi-step reasoning and coding in the Qwen ecosystem
Modalities: Text, reasoning, coding, tools and structured output
Context: See current Model Studio catalog
Published input: Verify
Published output: Verify
Caveat: Pricing, endpoint names and availability differ between China and international Model Studio regions.
Open official source (opens in a new tab)Best fit: Long-context agentic work and software engineering
Modalities: Text, software engineering, knowledge work, deep reasoning and tool calling
Context: 1M tokens
Published input: ¥20/1M
Published output: ¥100/1M
Caveat: Prices are official CNY rates per million tokens and must not be displayed as USD.
Open official source (opens in a new tab)Best fit: Long-horizon coding and autonomous agent workflows
Modalities: Text, reasoning, coding, long-running agents and tools
Context: 1M tokens
Published input: Verify
Published output: Verify
Caveat: Provider capability statements require independent task-specific reproduction.
Open official source (opens in a new tab)Best fit: Unified multimodal tasks and Chinese-language applications
Modalities: Unified text, image, video and audio understanding/generation
Context: 128K tokens
Published input: $1.4/1M
Published output: $5.6/1M
Caveat: The displayed API price is from Baidu’s international Qianfan catalog; China-region billing differs.
Open official source (opens in a new tab)Best fit: Production-oriented agent, coding and multimodal tasks
Modalities: General agents, coding and multimodal workflows
Context: See current Volcengine Ark catalog
Published input: Verify
Published output: Verify
Caveat: Volcengine pricing uses tiered regional tables; verify the exact model ID and input-length tier.
Open official source (opens in a new tab)Best fit: Agentic reasoning, coding and long-context work
Modalities: Text, multimodal chat input, coding, tool use and long-context agents
Context: 1M tokens
Published input: Verify
Published output: Verify
Caveat: Use the current regional pricing table and distinguish API pay-as-you-go from Token Plan subscriptions.
Open official source (opens in a new tab)Best fit: Low-cost multimodal reasoning and computer-use style tasks
Modalities: Multimodal reasoning, visual interaction, coding and agent workflows
Context: See current StepFun model catalog
Published input: ¥1.35/1M
Published output: ¥8.1/1M
Caveat: Official prices are CNY per million tokens. Confirm current quotas and model lifecycle notices.
Open official source (opens in a new tab)Best fit: Efficient mixture-of-experts reasoning in the Tencent ecosystem
Modalities: Text, hybrid reasoning, math, science, long documents and agents
Context: 224K max input; 32K max output
Published input: Verify
Published output: Verify
Caveat: Tencent is migrating newer model access toward TokenHub; verify the active platform before integration.
Open official source (opens in a new tab)Best fit: Dense layouts, small text, multilingual rendering, interfaces and infographics
Modalities: Text-to-image generation and image editing
Context: Up to 4.5K-token image instructions
Published input: Verify
Published output: Verify
Caveat: Capabilities and examples are provider-published; independent benchmark reproduction is pending.
Open official source (opens in a new tab)Best fit: Real-time Flash variant and higher-fidelity Plus variant
Modalities: Text-to-speech, voice cloning, style control and non-verbal tags
Context: 16-language text-to-speech release
Published input: Verify
Published output: Verify
Caveat: The 300ms-level first-packet latency and multilingual results are provider-reported; verify independently for production workloads.
Open official source (opens in a new tab)Use a small test set representing your real tasks. Do not choose from a single leaderboard score.
Include currency, output length, caching, tools, retries, batch discounts and long-context surcharges.
Review retention, regional processing, enterprise controls, export restrictions and whether a local model is practical.