{"auto_groups":["【GPT】plus/team炸弹车混合分组","【GPT】codex混合分组【不允许破限】","【GPT】百分百稳定分组，使用官方key作为兜底使用","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC可外接分组","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"data":[{"model_name":"Ali/deepseek-v4-pro","description":"[Alibaba Cloud Model Studio contract channel] DeepSeek-V4-Pro, workday peak/off-peak: Mon-Fri 09:00-12:00 \u0026 14:00-18:00 Beijing time = 70% of the official list price, ¥8.4 input / ¥0.7 cache hit / ¥16.8 output; all other hours and weekends = 50%, ¥6 / ¥0.5 / ¥12 per 1M tokens; group ratio 1.","vendor_id":7,"quota_type":0,"model_ratio":4.2,"model_price":0,"owner_by":"","completion_ratio":2,"cache_ratio":0.08333333333333333,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"weekday(\"UTC\") \u003e= 1 \u0026\u0026 weekday(\"UTC\") \u003c= 5 \u0026\u0026 ((hour(\"UTC\") \u003e= 1 \u0026\u0026 hour(\"UTC\") \u003c 4) || (hour(\"UTC\") \u003e= 6 \u0026\u0026 hour(\"UTC\") \u003c 10)) ? tier(\"peak\", p * 8.4 + c * 16.8 + cr * 0.7) : tier(\"off_peak\", p * 6 + c * 12 + cr * 0.5)","pricing_version":"5a90f2b86c08bd983a9a2e6d66c255f4eaef9c4bc934386d2b6ae84ef0ff1f1f"},{"model_name":"claude-sonnet-5-5","description":"Claude Sonnet 5.5 balances response speed and reasoning for coding, tool use and multi-step agents. It supports text/image input, 1M context, up to 128K output and adaptive thinking.","tags":"Reasoning,Tools,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"claude-opus-4.5","description":"A flagship Claude for demanding software engineering, agent workflows, and complex professional tasks. Strong Opus capability with a more efficient price point. This dotted name is a channel alias.","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"codex-auto-review","description":"A specialized Codex reviewer that checks proposed agent actions at escalation points. Built for risk assessment and approval decisions, not general-purpose chat.","vendor_id":1,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"]},{"model_name":"claude-haiku-4-5-20251001","description":"A pinned Haiku 4.5 snapshot for fast assistance, classification, and compact agent jobs. Small, responsive, and easy to scale.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":3,"quota_type":0,"model_ratio":0.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"glm-5.3","description":"GLM-5.3 is a text model for complex software engineering and sustained agent tasks, with 1M context and up to 128K output. The native API keeps thinking enabled with low, high and max effort, plus tools and structured output.","tags":"Reasoning,Tools,1M","vendor_id":8,"quota_type":0,"model_ratio":4,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"grok-4.5","description":"A current Grok for chat, code, and agent tools, tuned to keep hallucinations in check. A lively collaborator that should still verify important facts.","icon":"XAI.Color","tags":"Reasoning,Tools,Files,Vision,500K","vendor_id":6,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.15,"enable_groups":["【Grok】稳定分组"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 6 + cr * 0.3) : tier(\"long_context\", p * 4 + c * 12 + cr * 0.6)"},{"model_name":"gemini-3.8-flash","description":"Gemini 3.8 Flash is designed for extended software engineering, autonomous agents and complex enterprise workflows. Promotional token pricing runs through December 31, 2026.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.8-flash-tiered","description":"Gemini 3.8 Flash is designed for extended software engineering, autonomous agents and complex enterprise workflows. Promotional token pricing runs through December 31, 2026. tiered is a channel name; billing follows the prices shown here and does not identify a separate Google model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"ox-alpha","description":"Ox Alpha is the revealed preview name of Z.ai GLM-5.3-Flash, for coding, sustained agents and multimodal understanding with about 1M context. This gateway retains the alias; availability depends on the upstream preview endpoint.","tags":"Reasoning,Tools,Vision,1M","vendor_id":8,"quota_type":0,"model_ratio":0,"model_price":0,"owner_by":"","completion_ratio":1,"enable_groups":["【GPT】【不降智】可外接不降智"],"supported_endpoint_types":["openai"]},{"model_name":"claude-sonnet-4-5","description":"A pinned Sonnet 4.5 snapshot that balances coding, analysis, agent work, and spend. The dependable default when you need both quality and restraint. Undated Sonnet 4.5 name for coding and agent workflows.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":3,"quota_type":0,"model_ratio":1.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Kiro】自营分组"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"gpt-5.3-codex-spark","description":"Codex's ultra-fast, text-only research preview for real-time coding. Its 128K context and 1000+ token-per-second serving make targeted edits and tight iteration feel nearly instant.","vendor_id":1,"quota_type":0,"model_ratio":0.875,"model_price":0,"owner_by":"","completion_ratio":8,"cache_ratio":0.09999999999999999,"enable_groups":["【GPT】codex混合分组【不允许破限】"],"supported_endpoint_types":["openai"]},{"model_name":"claude-opus-4.7","description":"A stronger Opus tier for serious software work and decisions where shallow answers are expensive. Bring it the messy, high-stakes problems. This dotted name is a channel alias.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"claude-opus-4.8","description":"Top-shelf Opus for the hardest reasoning, coding, and long-running agents. Reserve it for work that deserves the extra thought. This dotted name is a channel alias.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"gemini-3.1-pro-preview","description":"Gemini 3.1 Pro Preview supports complex reasoning, multimodal understanding, coding and agents with long context. It is a preview model; prompts above 200K tokens use long-context pricing.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4)"},{"model_name":"gpt-5.6-sol","description":"The flagship GPT-5.6 for complex, open-ended, and high-value work. Choose Sol when difficult coding, deep research, computer use, cybersecurity, or polished professional output needs the family's full depth.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":1,"quota_type":0,"model_ratio":2,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 4 + c * 20 + cr * 0.4 + cc * 5) : tier(\"long_context\", p * 8 + c * 30 + cr * 0.8 + cc * 10)"},{"model_name":"claude-opus-4-5","description":"A flagship Claude for demanding software engineering, agent workflows, and complex professional tasks. Strong Opus capability with a more efficient price point.","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Kiro】自营分组"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + cr * 0.5 + cc * 6.25 + cc1h * 10 + c * 25)","credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"gpt-image-2.5-flare","description":"GPT Image 2.5 Flare focuses on fast, everyday image generation and editing from text/image inputs. Official pricing is token-based; this platform’s subscription relay uses the displayed fixed price, with quality and resolution support determined by the channel.","tags":"Vision,Image Generation","vendor_id":1,"quota_type":1,"model_ratio":0,"model_price":0.04,"owner_by":"","completion_ratio":0,"enable_groups":["【GPT】生图分组"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"Ali/glm-5.2","description":"[Alibaba Cloud Model Studio contract channel] GLM-5.2 at 50% of the official list price. ¥4 input / ¥1 cache hit / ¥14 output per 1M tokens; group ratio 1, the list price is what you pay.","vendor_id":8,"quota_type":0,"model_ratio":2,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"]},{"model_name":"Ali/kimi-k2.7-code","description":"[Alibaba Cloud Model Studio contract channel] Kimi-K2.7-Code at 70% of the official list price. ¥4.55 input / ¥0.91 cache hit / ¥18.9 output per 1M tokens (explicit cache read ¥0.455, create ¥5.6875); group ratio 1.","vendor_id":9,"quota_type":0,"model_ratio":2.275,"model_price":0,"owner_by":"","completion_ratio":4.153846153846154,"cache_ratio":0.2,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 4.55 + c * 18.9 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.455 : 0.91) + cc * 5.6875)"},{"model_name":"gemini-3.6-flash-high","description":"Gemini 3.6 Flash balances speed and multimodal capability for coding, information processing and tool use. This earlier Flash generation has promotional token pricing through December 31, 2026. high is a channel preset, not a separate official model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4.1-flash-0910","description":"DeepSeek V4.1 Flash is a multimodal mixture-of-experts model for coding, reasoning and tool use with 1M context. The official API recommends deepseek-flash; this gateway name and supported modalities depend on the channel. The configured channel maps this name to deepseek-v4.1-flash; it does not pin a separate official snapshot.","tags":"Reasoning,Tools,Vision,1M","vendor_id":7,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.02,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5) \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)) \u0026\u0026 !(now().Year() == 2026 \u0026\u0026 ((month(\"Asia/Shanghai\") == 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 3) || (month(\"Asia/Shanghai\") == 2 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 15 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 23) || (month(\"Asia/Shanghai\") == 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 6) || (month(\"Asia/Shanghai\") == 5 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 5) || (month(\"Asia/Shanghai\") == 6 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 19 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 21) || (month(\"Asia/Shanghai\") == 9 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 25 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 27) || (month(\"Asia/Shanghai\") == 10 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 7))) ? tier(\"peak\", p * 2 + c * 8 + cr * 0.04) : tier(\"off_peak\", p * 1 + c * 4 + cr * 0.02)"},{"model_name":"gpt-6-astra","description":"GPT-6 Astra handles demanding reasoning, coding, research, computer use and documents. It supports text/image input, about 1.05M context and up to 128K output, with low-to-max reasoning and higher pricing above 272K input tokens.","tags":"Reasoning,Tools,Vision,1.05M","vendor_id":1,"quota_type":0,"model_ratio":5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 10 + c * 50 + cr * 1 + cc * 12.5) : tier(\"long_context\", p * 20 + c * 75 + cr * 2 + cc * 25)"},{"model_name":"claude-haiku-4-5","description":"The quick, economical Claude lane: ideal for chat, sorting, summaries, and lightweight agents that need to stay snappy.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":3,"quota_type":0,"model_ratio":0.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"claude-opus-4-8","description":"Top-shelf Opus for the hardest reasoning, coding, and long-running agents. Reserve it for work that deserves the extra thought.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + cr * 0.5 + cc * 6.25 + cc1h * 10 + c * 25)","credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"claude-fable-5","description":"Claude Fable 5 is a general model for demanding reasoning, research and extended agent work, including deep analysis and multi-step planning; it is not a creative-writing-only model.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"claude-fable-5-1","description":"Claude Fable 5.1 supports demanding reasoning and extended agent workflows, with text/image input, 1M context and up to 128K output. Adaptive thinking is always enabled.","tags":"Reasoning,Tools,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.025,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"claude-opus-4.6","description":"Premium Claude for difficult code, careful plans, and expert-level reasoning. It takes its time, but usually uses that time well. This dotted name is a channel alias.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + c * 25 + cr * 0.5 + cc * 6.25 + cc1h * 10)"},{"model_name":"claude-opus-5.5","description":"Claude Opus 5.5 is designed for sustained agentic coding and knowledge work. It supports text/image input, 1M context, up to 128K output and always-on adaptive thinking. This dotted name is a channel alias; available features depend on the serving channel.","tags":"Reasoning,Tools,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.05,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"deepseek-v4-pro","description":"The V4 flagship: 1M-token context plus a serious thinking mode, tuned for complex code, deep analysis, and long-horizon agents.","vendor_id":7,"quota_type":0,"model_ratio":4.5,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.03333333333333333,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5) \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)) \u0026\u0026 !(now().Year() == 2026 \u0026\u0026 ((month(\"Asia/Shanghai\") == 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 3) || (month(\"Asia/Shanghai\") == 2 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 15 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 23) || (month(\"Asia/Shanghai\") == 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 6) || (month(\"Asia/Shanghai\") == 5 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 5) || (month(\"Asia/Shanghai\") == 6 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 19 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 21) || (month(\"Asia/Shanghai\") == 9 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 25 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 27) || (month(\"Asia/Shanghai\") == 10 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 7))) ? tier(\"peak\", p * 9 + c * 27 + cr * 0.3) : tier(\"off_peak\", p * 4.5 + c * 13.5 + cr * 0.15)"},{"model_name":"gpt-5.5","description":"The frontier GPT to start with for coding, computer use, research, and knowledge work. Broadly capable enough to be the default without feeling bland.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":1,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 5 + c * 30 + cr * 0.5) : tier(\"long_context\", p * 10 + c * 45 + cr * 1)"},{"model_name":"claude-sonnet-5","description":"A balanced Claude for code, analysis, and agent workflows, with an eye on the bill. A sensible everyday choice before reaching for Opus.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"gpt-image-2","description":"The visual specialist for generating and editing images from prompts. Great for turning a rough idea into something you can actually look at and refine.","icon":"OpenAI.Color","tags":"Files,Vision","vendor_id":1,"quota_type":1,"model_ratio":0,"model_price":0.05,"owner_by":"","completion_ratio":0,"enable_groups":["【GPT】生图分组"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"Ali/kimi-k3","description":"[Alibaba Cloud Model Studio contract channel] Kimi-K3 at 95% of the official list price. ¥19 input / ¥1.9 cache hit / ¥95 output per 1M tokens; group ratio 1.","vendor_id":9,"quota_type":0,"model_ratio":9.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.1-pro-low","description":"Gemini 3.1 Pro Preview supports complex reasoning, multimodal understanding, coding and agents with long context. It is a preview model; prompts above 200K tokens use long-context pricing. low is a channel preset suffix; actual thinking settings depend on the channel.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4)"},{"model_name":"Ali/kimi-k2.6","description":"[Alibaba Cloud Model Studio contract channel] Kimi-K2.6 at 70% of the official list price. ¥4.55 input / ¥0.91 cache hit / ¥18.9 output per 1M tokens (explicit cache read ¥0.455, create ¥5.6875); group ratio 1.","vendor_id":9,"quota_type":0,"model_ratio":2.275,"model_price":0,"owner_by":"","completion_ratio":4.153846153846154,"cache_ratio":0.2,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 4.55 + c * 18.9 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.455 : 0.91) + cc * 5.6875)"},{"model_name":"claude-opus-4-5-20251101","description":"A flagship Claude snapshot for thorny reasoning, serious coding, and agents that need to keep the thread over long runs.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"grok-4.7","description":"Grok 4.7 supports coding, agents and knowledge work with text/image input, 500K context, function calling and structured output. Reasoning levels range from low to xhigh; current information requires search tools enabled by the channel.","tags":"Reasoning,Tools,Vision,500K","vendor_id":6,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["【Grok】稳定分组"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 6 + cr * 0.5) : tier(\"long_context\", p * 4 + c * 12 + cr * 1)"},{"model_name":"Ali/qwen3.7-max","description":"[Alibaba Cloud Model Studio contract channel] Qwen3.7-Max at 60% of the official list price. ¥7.2 input / ¥1.44 cache hit / ¥21.6 output per 1M tokens (explicit cache read ¥0.72, create ¥9); group ratio 1.","vendor_id":16,"quota_type":0,"model_ratio":3.6,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.2,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 7.2 + c * 21.6 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.72 : 1.44) + cc * 9)"},{"model_name":"deepseek-v4-flash-0731","description":"A pinned V4 Flash snapshot: same 1M context and snappy everyday feel, frozen so your pipeline runs exactly the same tomorrow.","vendor_id":7,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.02,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5) \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)) \u0026\u0026 !(now().Year() == 2026 \u0026\u0026 ((month(\"Asia/Shanghai\") == 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 3) || (month(\"Asia/Shanghai\") == 2 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 15 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 23) || (month(\"Asia/Shanghai\") == 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 6) || (month(\"Asia/Shanghai\") == 5 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 5) || (month(\"Asia/Shanghai\") == 6 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 19 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 21) || (month(\"Asia/Shanghai\") == 9 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 25 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 27) || (month(\"Asia/Shanghai\") == 10 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 7))) ? tier(\"peak\", p * 2 + c * 8 + cr * 0.04) : tier(\"off_peak\", p * 1 + c * 4 + cr * 0.02)"},{"model_name":"Ali/qwen3.8-flash","description":"[Alibaba Cloud Model Studio contract channel] Qwen3.8-Flash at 85% of the official list price. ¥0.68 input / ¥0.085 cache hit / ¥2.295 output per 1M tokens (explicit cache create ¥1.0625); group ratio 1. Fast, low-cost Qwen 3.8 for high-volume everyday tasks.","vendor_id":16,"quota_type":0,"model_ratio":0.34,"model_price":0,"owner_by":"","completion_ratio":3.375,"cache_ratio":0.125,"create_cache_ratio":1.5625,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 0.68 + c * 2.295 + cr * 0.085 + cc * 1.0625)"},{"model_name":"gpt-6-sol","description":"GPT-6 Sol supports complex coding and agent workflows with text/image input and about 1.05M context. Use Responses for built-in tools; Chat Completions function calling requires reasoning_effort=none.","tags":"Reasoning,Tools,Vision,1.05M","vendor_id":1,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 10 + cr * 0.2 + cc * 2.5) : tier(\"long_context\", p * 4 + c * 15 + cr * 0.4 + cc * 5)"},{"model_name":"Ali/deepseek-v4-pro-0813","description":"[Alibaba Cloud Model Studio contract channel] DeepSeek-V4-Pro-0813, workday peak/off-peak: Mon-Fri 09:00-12:00 \u0026 14:00-18:00 Beijing time = 70% of the official list price, ¥6.3 input / ¥0.21 cache hit / ¥18.9 output; all other hours and weekends = the official off-peak price, ¥4.5 / ¥0.15 / ¥13.5 per 1M tokens; group ratio 1.","vendor_id":7,"quota_type":0,"model_ratio":3.15,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.03333333333333333,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"weekday(\"UTC\") \u003e= 1 \u0026\u0026 weekday(\"UTC\") \u003c= 5 \u0026\u0026 ((hour(\"UTC\") \u003e= 1 \u0026\u0026 hour(\"UTC\") \u003c 4) || (hour(\"UTC\") \u003e= 6 \u0026\u0026 hour(\"UTC\") \u003c 10)) ? tier(\"peak\", p * 6.3 + c * 18.9 + cr * 0.21) : tier(\"off_peak\", p * 4.5 + c * 13.5 + cr * 0.15)"},{"model_name":"Ali/deepseek-v4-flash-0731","description":"[Alibaba Cloud Model Studio contract channel] DeepSeek-V4-Flash-0731 at 70% of the official list price, billed by workday peak/off-peak: Mon-Fri 09:00-12:00 \u0026 14:00-18:00 Beijing time ¥2.1 input / ¥0.21 cache hit / ¥6.3 output; all other hours and weekends ¥1.05 / ¥0.105 / ¥3.15 per 1M tokens; group ratio 1.","vendor_id":7,"quota_type":0,"model_ratio":1.05,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.1,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"weekday(\"UTC\") \u003e= 1 \u0026\u0026 weekday(\"UTC\") \u003c= 5 \u0026\u0026 ((hour(\"UTC\") \u003e= 1 \u0026\u0026 hour(\"UTC\") \u003c 4) || (hour(\"UTC\") \u003e= 6 \u0026\u0026 hour(\"UTC\") \u003c 10)) ? tier(\"peak\", p * 2.1 + c * 6.3 + cr * 0.21) : tier(\"off_peak\", p * 1.05 + c * 3.15 + cr * 0.105)"},{"model_name":"gemini-3.6-flash-medium","description":"Gemini 3.6 Flash balances speed and multimodal capability for coding, information processing and tool use. This earlier Flash generation has promotional token pricing through December 31, 2026. medium is a channel preset, not a separate official model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"Ali/qwen3.8-max","description":"[Alibaba Cloud Model Studio contract channel] Qwen3.8-Max at 85% of the official list price. ¥10.2 input / ¥1.275 cache hit / ¥30.6 output per 1M tokens (explicit cache read ¥0.85, create ¥12.75); group ratio 1. Alibaba multimodal flagship: 1M context, long-video understanding, agent skills.","vendor_id":16,"quota_type":0,"model_ratio":5.1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.125,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"base\", p * 10.2 + c * 30.6 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.85 : 1.275) + cc * 12.75)"},{"model_name":"gemini-3.6-flash","description":"Gemini 3.6 Flash balances speed and multimodal capability for coding, information processing and tool use. This earlier Flash generation has promotional token pricing through December 31, 2026.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.1-pro","description":"Gemini 3.1 Pro Preview supports complex reasoning, multimodal understanding, coding and agents with long context. It is a preview model; prompts above 200K tokens use long-context pricing. This channel name omits preview; the name alone does not establish a separate stable release.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4)"},{"model_name":"gpt-5.6-terra","description":"The pragmatic GPT-5.6 all-rounder, balancing intelligence and cost. A natural everyday choice for strong reasoning, coding, writing, and tool use when Sol's full depth is unnecessary.","tags":"Reasoning,Tools,Files,Vision,1.1M","vendor_id":1,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2 + cc * 2.5) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4 + cc * 5)"},{"model_name":"claude-opus-4-7","description":"A stronger Opus tier for serious software work and decisions where shallow answers are expensive. Bring it the messy, high-stakes problems.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + cr * 0.5 + cc * 6.25 + cc1h * 10 + c * 25)","credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"gpt-image-2.5","description":"This is a channel-provided image generation/editing name. Some configured channels map it to gpt-image-2-auto, so the name does not guarantee GPT Image 2.5 Flare or Sunburst. The displayed fixed price applies.","tags":"Image Generation","vendor_id":1,"quota_type":1,"model_ratio":0,"model_price":0.05,"owner_by":"","completion_ratio":0,"enable_groups":["【GPT】生图分组"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"gemini-3-flash","description":"Gemini 3 Flash combines multimodal understanding with responsive reasoning for chat, coding and agents. This is a channel name; the documented preview model is gemini-3-flash-preview.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.25,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.1-flash-image","description":"Gemini 3.1 Flash Image (Nano Banana 2) generates and edits images from text and image inputs for interactive and high-throughput creation. Text and image output tokens have separate prices.","tags":"Vision,Image Generation","vendor_id":21,"quota_type":0,"model_ratio":0.25,"model_price":0,"owner_by":"","completion_ratio":6,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 0.5 + c * 3 + img_o * 60)"},{"model_name":"claude-sonnet-4-6","description":"The Claude workhorse for coding agents, careful analysis, and keeping production costs sane. Not flashy; reliably useful.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":1.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"claude-opus-5","description":"Flagship Claude for deep reasoning, demanding code, and agents with a long memory of the job. The heavyweight when the task is not routine.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"gemini-3.7-flash-tiered","description":"Gemini 3.7 Flash focuses on efficient coding, agentic tool use and reliable multi-step execution. Promotional token pricing runs through December 31, 2026. tiered is a channel name; billing follows the prices shown here and does not identify a separate Google model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"glm-5.1","description":"GLM-5.1 is a text model from Zhipu for sustained tasks, reasoning, long documents, code and tool use. It offers about 200K context and up to 128K output, with a separate price tier above 32K input tokens.","tags":"Reasoning,Tools,200K","vendor_id":8,"quota_type":0,"model_ratio":3,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.19999999999999998,"create_cache_ratio":1.25,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 32000 ? tier(\"0-32K\", p * 6 + c * 24 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.6 : 1.2) + cc * 7.5) : tier(\"32K-200K\", p * 8 + c * 28 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.8 : 1.6) + cc * 10)"},{"model_name":"gemini-3.6-flash-low","description":"Gemini 3.6 Flash balances speed and multimodal capability for coding, information processing and tool use. This earlier Flash generation has promotional token pricing through December 31, 2026. low is a channel preset, not a separate official model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"Ali/qwen3.7-plus","description":"[Alibaba Cloud Model Studio contract channel] Qwen3.7-Plus at 60% of the official list price. 0-256K: ¥1.2 input / ¥0.24 cache hit / ¥4.8 output; 256K-1M: ¥3.6 / ¥0.72 / ¥14.4 per 1M tokens (explicit cache read ¥0.12, create ¥1.5); group ratio 1.","vendor_id":16,"quota_type":0,"model_ratio":0.6,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 256000 ? tier(\"0-256K\", p * 1.2 + c * 4.8 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.12 : 0.24) + cc * 1.5) : tier(\"256K-1M\", p * 3.6 + c * 14.4 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.36 : 0.72) + cc * 4.5)"},{"model_name":"grok-4.6","description":"xAI's frontier Grok for coding, agentic tasks, and knowledge work. It combines a 500K context, image understanding, strong tools, and configurable reasoning up to xhigh for demanding jobs.","vendor_id":6,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":3,"cache_ratio":0.25,"enable_groups":["【Grok】稳定分组"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 6 + cr * 0.5) : tier(\"long_context\", p * 4 + c * 12 + cr * 1)"},{"model_name":"gemini-3.7-flash","description":"Gemini 3.7 Flash focuses on efficient coding, agentic tool use and reliable multi-step execution. Promotional token pricing runs through December 31, 2026.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"gemini-3.1-pro-high","description":"Gemini 3.1 Pro Preview supports complex reasoning, multimodal understanding, coding and agents with long context. It is a preview model; prompts above 200K tokens use long-context pricing. high is a channel preset suffix; actual thinking settings depend on the channel.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":6,"cache_ratio":0.1,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 200000 ? tier(\"standard\", p * 2 + c * 12 + cr * 0.2) : tier(\"long_context\", p * 4 + c * 18 + cr * 0.4)"},{"model_name":"deepseek-v4-flash","description":"The quick, affordable V4 lane: a 1M-token context, a thinking mode you can switch on or off, and plenty of brain for everyday work at a friendly price.","vendor_id":7,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.02,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5) \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)) \u0026\u0026 !(now().Year() == 2026 \u0026\u0026 ((month(\"Asia/Shanghai\") == 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 3) || (month(\"Asia/Shanghai\") == 2 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 15 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 23) || (month(\"Asia/Shanghai\") == 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 6) || (month(\"Asia/Shanghai\") == 5 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 5) || (month(\"Asia/Shanghai\") == 6 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 19 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 21) || (month(\"Asia/Shanghai\") == 9 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 25 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 27) || (month(\"Asia/Shanghai\") == 10 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 7))) ? tier(\"peak\", p * 2 + c * 8 + cr * 0.04) : tier(\"off_peak\", p * 1 + c * 4 + cr * 0.02)"},{"model_name":"kimi-k3","description":"Moonshot's 2.8-trillion-parameter flagship: multimodal understanding, 1M context, and the stamina to engineer for hours with little supervision.","vendor_id":9,"quota_type":0,"model_ratio":10,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 20 + c * 100 + cr * 2 + cc * 20 + cc1h * 40)"},{"model_name":"gpt-image-2.5-sunburst","description":"GPT Image 2.5 Sunburst focuses on capable image generation and precise editing. Official pricing is token-based; this platform uses the displayed fixed price, and supported parameters depend on the channel.","tags":"Vision,Image Generation","vendor_id":1,"quota_type":1,"model_ratio":0,"model_price":0.06,"owner_by":"","completion_ratio":0,"enable_groups":["【GPT】生图分组"],"supported_endpoint_types":["image-generation","openai"]},{"model_name":"gemini-3.6-flash-tiered","description":"Gemini 3.6 Flash balances speed and multimodal capability for coding, information processing and tool use. This earlier Flash generation has promotional token pricing through December 31, 2026. tiered is a channel name; billing follows the prices shown here and does not identify a separate Google model.","tags":"Reasoning,Tools,Vision","vendor_id":21,"quota_type":0,"model_ratio":0.375,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.09999999999999999,"enable_groups":["【gemini】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"deepseek-v4.1-flash","description":"DeepSeek V4.1 Flash is a multimodal mixture-of-experts model for coding, reasoning and tool use with 1M context. The official API recommends deepseek-flash; this gateway name and supported modalities depend on the channel.","tags":"Reasoning,Tools,Vision,1M","vendor_id":7,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.02,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"(weekday(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 weekday(\"Asia/Shanghai\") \u003c= 5) \u0026\u0026 ((hour(\"Asia/Shanghai\") \u003e= 9 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 12) || (hour(\"Asia/Shanghai\") \u003e= 14 \u0026\u0026 hour(\"Asia/Shanghai\") \u003c 18)) \u0026\u0026 !(now().Year() == 2026 \u0026\u0026 ((month(\"Asia/Shanghai\") == 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 3) || (month(\"Asia/Shanghai\") == 2 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 15 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 23) || (month(\"Asia/Shanghai\") == 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 4 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 6) || (month(\"Asia/Shanghai\") == 5 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 5) || (month(\"Asia/Shanghai\") == 6 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 19 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 21) || (month(\"Asia/Shanghai\") == 9 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 25 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 27) || (month(\"Asia/Shanghai\") == 10 \u0026\u0026 day(\"Asia/Shanghai\") \u003e= 1 \u0026\u0026 day(\"Asia/Shanghai\") \u003c= 7))) ? tier(\"peak\", p * 2 + c * 8 + cr * 0.04) : tier(\"off_peak\", p * 1 + c * 4 + cr * 0.02)"},{"model_name":"glm-5.3-flash","description":"GLM-5.3-Flash is Z.ai’s native multimodal model for coding, tools and professional documents. It supports text, images, video and files, uses 320B total/18B active parameters, and was previewed as Ox Alpha.","tags":"Reasoning,Tools,Vision,1M","vendor_id":8,"quota_type":0,"model_ratio":0.4,"model_price":0,"owner_by":"","completion_ratio":3.4999999999999996,"cache_ratio":0.2875,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"claude-opus-4-6","description":"Premium Claude for difficult code, careful plans, and expert-level reasoning. It takes its time, but usually uses that time well.","tags":"Reasoning,Tools,Files,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"billing_mode":"tiered_expr","billing_expr":"tier(\"standard\", p * 5 + cr * 0.5 + cc * 6.25 + cc1h * 10 + c * 25)","credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"claude-sonnet-4-5-20250929","description":"A pinned Sonnet 4.5 snapshot that balances coding, analysis, agent work, and spend. The dependable default when you need both quality and restraint.","tags":"Reasoning,Tools,Files,Vision,200K","vendor_id":3,"quota_type":0,"model_ratio":1.5,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.1,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】CC可外接分组"],"supported_endpoint_types":["anthropic","openai"]},{"model_name":"claude-opus-5-5","description":"Claude Opus 5.5 is designed for sustained agentic coding and knowledge work. It supports text/image input, 1M context, up to 128K output and always-on adaptive thinking.","tags":"Reasoning,Tools,Vision,1M","vendor_id":3,"quota_type":0,"model_ratio":2,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.05,"create_cache_ratio":1.25,"enable_groups":["【Claude】【非自营】kiro高智分组","【Claude】【非自营】可外接，蒸馏","【Kiro】自营分组","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC纯享分组（可外接，蒸馏）"],"supported_endpoint_types":["anthropic","openai"],"credit_pricing":[{"group":"【Kiro】自营分组","mode":"charge","unit":"kiro_credit","currency":"CNY","price_per_credit":"0.11","group_ratio":"1","ratio_source":"group","effective_price_per_credit":"0.11"}]},{"model_name":"Ali/glm-5.1","description":"[Alibaba Cloud Model Studio contract channel] GLM-5.1 at 70% of the official list price. 0-32K: ¥4.2 input / ¥0.84 cache hit / ¥16.8 output; 32K-200K: ¥5.6 / ¥1.12 / ¥19.6 per 1M tokens; group ratio 1.","vendor_id":8,"quota_type":0,"model_ratio":2.1,"model_price":0,"owner_by":"","completion_ratio":4,"cache_ratio":0.2,"create_cache_ratio":1.25,"enable_groups":["【阿里云百炼折扣】"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 32000 ? tier(\"0-32K\", p * 4.2 + c * 16.8 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.42 : 0.84) + cc * 5.25) : tier(\"32K-200K\", p * 5.6 + c * 19.6 + cr * (has(param(\"messages\"), \"cache_control\") || has(param(\"system\"), \"cache_control\") ? 0.56 : 1.12) + cc * 7)"},{"model_name":"glm-5.2","description":"Zhipu's strongest open-source flagship, built for long jobs: a genuinely usable 1M context, stable multi-hour runs, and top-tier coding.","vendor_id":8,"quota_type":0,"model_ratio":4,"model_price":0,"owner_by":"","completion_ratio":3.5,"cache_ratio":0.25,"enable_groups":["【短期国模优惠渠道】【非自营】"],"supported_endpoint_types":["openai"]},{"model_name":"gpt-6.1-sol","description":"GPT-6.1 Sol balances cost and capability for complex coding, computer use and professional work, with about 1.05M context. Tool calling uses Responses; reasoning supports low through max, but not none.","tags":"Reasoning,Tools,Vision,1.05M","vendor_id":1,"quota_type":0,"model_ratio":1,"model_price":0,"owner_by":"","completion_ratio":5,"cache_ratio":0.05,"create_cache_ratio":1.25,"enable_groups":["【GPT】plus/team炸弹车混合分组","【GPT】plus分组","【GPT】codex混合分组【不允许破限】","【GPT】codexPro分组【不允许破限】","【GPT】【不降智】可外接不降智","【GPT】仅限coding分组","【GPT】百分百稳定分组，使用官方key作为兜底使用"],"supported_endpoint_types":["openai"],"billing_mode":"tiered_expr","billing_expr":"len \u003c= 272000 ? tier(\"standard\", p * 2 + c * 10 + cr * 0.1 + cc * 2.5) : tier(\"long_context\", p * 4 + c * 15 + cr * 0.2 + cc * 5)"}],"group_names":{"auto":"auto","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】":"【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】","【Claude】【非自营】CC可外接分组":"【Claude】【非自营】CC可外接分组","【Claude】【非自营】CC纯享分组（可外接，蒸馏）":"【Claude】【非自营】CC纯享分组（可外接，蒸馏）","【Claude】【非自营】kiro高智分组":"【Claude】【非自营】kiro高智分组","【Claude】【非自营】可外接，蒸馏":"【Claude】【非自营】可外接，蒸馏","【GPT】codexPro分组【不允许破限】":"【GPT】codexPro分组【不允许破限】","【GPT】codex混合分组【不允许破限】":"【GPT】codex混合分组【不允许破限】","【GPT】plus/team炸弹车混合分组":"【GPT】plus/team炸弹车混合分组","【GPT】plus分组":"【GPT】plus分组","【GPT】【不降智】可外接不降智":"【GPT】【不降智】可外接不降智","【GPT】仅限coding分组":"【GPT】仅限coding分组","【GPT】生图分组":"【GPT】生图分组","【GPT】百分百稳定分组，使用官方key作为兜底使用":"【GPT】百分百稳定分组，使用官方key作为兜底使用","【Grok】稳定分组":"【Grok】稳定分组","【Kiro】自营分组":"【Kiro】自营分组","【gemini】【非自营】":"【gemini】【非自营】","【短期国模优惠渠道】【非自营】":"【短期国模优惠渠道】【非自营】","【阿里云百炼折扣】":"【阿里云百炼折扣】"},"group_ratio":{"auto":1,"【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】":1.15,"【Claude】【非自营】CC可外接分组":0.9,"【Claude】【非自营】CC纯享分组（可外接，蒸馏）":1.35,"【Claude】【非自营】kiro高智分组":0.4,"【Claude】【非自营】可外接，蒸馏":1,"【GPT】codexPro分组【不允许破限】":0.2,"【GPT】codex混合分组【不允许破限】":0.17,"【GPT】plus/team炸弹车混合分组":0.08,"【GPT】plus分组":0.12,"【GPT】【不降智】可外接不降智":0.3,"【GPT】仅限coding分组":0.4,"【GPT】生图分组":1,"【GPT】百分百稳定分组，使用官方key作为兜底使用":0.6,"【Grok】稳定分组":0.2,"【Kiro】自营分组":1,"【gemini】【非自营】":0.2,"【短期国模优惠渠道】【非自营】":0.25,"【阿里云百炼折扣】":1},"pricing_version":"a42d372ccf0b5dd13ecf71203521f9d2","recommended_model_cards":[{"id":"card-mr4iyssy-01","model_name":"claude-fable-5-1","title":"集A/大成之作","image_url":"https://i.imgs.ovh/2026/09/02/04a2c6481389e26ee841ac146fe8c326.png","description":"A/骂归骂，但是模型还是独一档的","enabled":true,"sort":1},{"id":"card-mrnq96r2-03","model_name":"grok-4.6","title":"Xai最新力作","image_url":"https://i.imgs.ovh/2026/09/02/6527faf7897060e78fa29f80a8d3cea6.png","description":"老马模型还可以，迭代快","enabled":true,"sort":2},{"id":"card-mrr6z9vv-04","model_name":"kimi-k3","title":"国模崛起之作","image_url":"https://i.imgs.ovh/2026/07/19/9beb733d0347523c2144ca9951bc814e.png","description":"国模目前集大成之作！kimi你拯救了一切！缺点是价格和肥波对标了...","enabled":true,"sort":2},{"id":"card-mr4je7vo-02","model_name":"gpt-5.6-sol","title":"OAI在交出的答卷","image_url":"https://i.imgs.ovh/2026/07/17/ddd4963cc7d85cfdbc2e8436803810f2.png","description":"思考可能会很长，很喜欢调用智能体，建议开高即可","enabled":true,"sort":3},{"id":"card-mtk26rx5-05","model_name":"qwen3.8-max","title":"QWEN","image_url":"https://i.imgs.ovh/2026/09/02/f26869a5a98ac52cba4664392ddf0230.png","description":"qwen刷分王，训练集超高分，网安有奇效","enabled":true,"sort":20}],"recommended_model_cards_autoplay_enabled":true,"recommended_model_cards_interval_seconds":2,"success":true,"supported_endpoint":{"anthropic":{"path":"/v1/messages","method":"POST"},"image-generation":{"path":"/v1/images/generations","method":"POST"},"openai":{"path":"/v1/chat/completions","method":"POST"}},"usable_group":{"auto":"自动分组：自动选择最低价可用分组，故障自动切换，可能会存在部分问题，推荐使用分组链来优化","【Claude】【自营】仅限Claude code使用，禁止外接【不允许破限】":"自营CC分组，单渠道，保真保纯，仅限claudecode，不允许破限，需要补号，稳定性不高","【Claude】【非自营】CC可外接分组":"非自营，可能存在掺水或者其他情况，接入前需要自行检测，无法保障","【Claude】【非自营】CC纯享分组（可外接，蒸馏）":"非自营，稳定性高，号池全部验证过了，支持直接跟我反馈上游是否降智，接入前需要自行检测","【Claude】【非自营】kiro高智分组":"kiro付费pro，智商在线，缓存需要查看下官方情况","【Claude】【非自营】可外接，蒸馏":"非自营，稳定性高，号池全部验证过了，支持直接跟我反馈上游是否降智，接入前需要自行检测","【GPT】codexPro分组【不允许破限】":"降智情况看检测，尽量保障","【GPT】codex混合分组【不允许破限】":"Pro号，以及预备升级pro的plus号组成（在升级pro前会先用plus测试下并发是否会被封号，并且有升级的过程，oai更少封号）","【GPT】plus/team炸弹车混合分组":"本分组不保证可用性，缓存高低纯看运气","【GPT】plus分组":"实际上现在是team分组了，plus价格很贵了，降智情况看检测，有的时候还可以","【GPT】【不降智】可外接不降智":"可以外接，但是无法保障缓存，不降智","【GPT】仅限coding分组":"本分组会对非coding请求限制，蒸馏、数据、翻译、破限等其他情况，会直接将用户禁用该分组，建议使用codex，其他cli编程工具不一定完美适配缓存情况。","【GPT】生图分组":"codex订阅转发生图，实际分辨率可能与参数不一致！","【GPT】百分百稳定分组，使用官方key作为兜底使用":"百分百稳定分组，使用官方key作为兜底使用，pro号池独立，本分组缓存低主要是测试用户比较多，发一个hi就不用了","【Grok】稳定分组":"grok的付费分组，非免费号池","【Kiro】自营分组":"kiro自营，智商高，缓存测是80-90区间，计费方式是积分制了，跟随kiro积分扣费","【gemini】【非自营】":"gemini模型渠道，尝鲜测试，效果好就转自营","【短期国模优惠渠道】【非自营】":"可用性自测，模型为真，缓存建议去性能页看一下再用","【阿里云百炼折扣】":"阿里云百炼官方合同价分组（倍率 1）：模型名前缀 Ali/，折扣已直接写进模型单价，各模型相对官方价格的折扣见模型描述；DeepSeek 按工作日峰谷计价，周末全天谷价"},"vendors":[{"id":21,"name":"Google","icon":"Gemini.Color"},{"id":11,"name":"Vivgrid"},{"id":14,"name":"UnoRouter"},{"id":7,"name":"DeepSeek","icon":"DeepSeek.Color"},{"id":10,"name":"Snowflake Cortex"},{"id":15,"name":"Xpersona"},{"id":16,"name":"阿里巴巴","icon":"Qwen.Color"},{"id":19,"name":"Meta","icon":"Meta"},{"id":22,"name":"MiniMax","icon":"Minimax.Color"},{"id":1,"name":"OpenAI","icon":"OpenAI"},{"id":2,"name":"讯飞","icon":"Spark.Color"},{"id":8,"name":"智谱","icon":"Zhipu.Color"},{"id":9,"name":"Moonshot","icon":"Moonshot"},{"id":13,"name":"Venice AI","icon":"VeniceAI"},{"id":3,"name":"Anthropic","icon":"Claude.Color"},{"id":17,"name":"腾讯","icon":"Hunyuan.Color"},{"id":18,"name":"小米","icon":"XiaomiMiMo"},{"id":6,"name":"xAI","icon":"XAI"},{"id":12,"name":"QiHang"},{"id":20,"name":"Stealth"}]}