1.05M-Token Context
Keep repository files, specifications and earlier findings in one request. The model supports a 1,050,000-token context window and up to 128,000 output tokens; large prompts still need retrieval discipline and a cost check.
GPT 6 Sol
Access GPT-6 Sol on GPTProto for demanding coding, code review and multi-step agent work. Use its 1.05M-token context and adjustable reasoning to handle substantial task material. Check the live rate and try the model with a GPTProto API key.
API 呼叫範例
curl --request POST "https://gptproto.com/v1/chat/completions" \
--header "Authorization: Bearer $GPTPROTO_API_KEY" \
--header "Content-Type: application/json" \
--data '{
"model": "gpt-6-sol",
"messages": [
{
"role": "user",
"content": "Hello"
}
]
}'GPT 6 Sol pricing
Chat, coding agents & document work. Priced per 1M tokens — input, cached input and output are billed separately. GPTProto is 20% below official rates.
你的用量
OpenAI · ≈ 148M tokens/mo（48M 快取）
OpenRouter
牌價 + 5.5% 儲值手續費
$288.65
每月
輸入（未快取）$67.52
快取讀取$10.128
輸出$211
OpenAI
直接向 OpenAI 購買（牌價）
$273.60
每月
輸入（未快取）$64
快取讀取$9.6
輸出$200
最划算
GPTProto
官方價 20% off
$218.88
每月
折扣後$218.88
實際成本$218.88
各來源月成本
OpenRouter
$288.65
OpenAI
$273.60
GPTProto
$218.88
每月節省 $54.72
在 GPTProto 相較 OpenAI（−20%）
年省約 $656.64 · 隨用隨付
OpenRouter 成本含約 5.5% 儲值手續費。GPTProto 在官方價上另有模型折扣（10–30% off），贈送額度也按折扣價消耗——節省會疊加。預估假設 60% 快取命中率。
相關模型
所有模型
| 模型 | 輸入 → 輸出 | ||||
|---|---|---|---|---|---|
| 1.05M | $1.60 / $8.00 每 1M | $2.00 / $0.16 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | |
| 1.05M | $0.08 / $0.40 每 1M | $0.10 / $0.008 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| — | $0.03 / — 每 1M | — | 輸入: 文字 輸出: 文字 | ||
| 500K | $1.20 / $3.60 每 1M | — / $0.30 每 1M | 輸入: 文字輸入: 圖像 輸出: 文字 | ||
| — | — | $0.30 / $1.20 每 1M | — / $0.006 每 1M | 輸入: 文字輸入: 圖像 輸出: 文字 | |
| 1M | $0.79 / $2.38 每 1M | — / $0.04 每 1M | 輸入: 文字 輸出: 文字 | ||
| 1.05M | $8.00 / $40.00 每 1M | $10.00 / $0.80 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $0.90 / $4.50 每 1M | $0.60 / $0.09 每 1M | 輸入: 文字輸入: 圖像輸入: 影片輸入: 文件輸入: 音訊 輸出: 文字 | ||
| 1M | $9.00 / $45.00 每 1M | $11.25 / $0.23 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1M | $1.80 / $5.40 每 1M | $2.25 / $0.23 每 1M | 輸入: 文字輸入: 圖像輸入: 影片輸入: 文件輸入: 音訊 輸出: 文字 | ||
| — | 1.31M | $0.15 / $0.50 每 1M | — / $0.03 每 1M | 輸入: 文字輸入: 圖像輸入: 影片輸入: 文件 輸出: 文字 | |
| — | 1.05M | $0.30 / $1.20 每 1M | — / $0.006 每 1M | 輸入: 文字輸入: 圖像 輸出: 文字 | |
| 1.31M | $1.26 / $3.96 每 1M | — / $0.23 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $0.90 / $4.50 每 1M | $0.60 / $0.09 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 500K | $1.20 / $3.60 每 1M | — / $0.30 每 1M | 輸入: 文字輸入: 圖像 輸出: 文字 | ||
| 1M | $1.80 / $5.40 每 1M | $2.25 / $0.23 每 1M | 輸入: 文字輸入: 圖像輸入: 影片輸入: 文件 輸出: 文字 | ||
| 1M | $4.50 / $22.50 每 1M | $5.63 / $0.45 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $0.90 / $4.50 每 1M | $0.60 / $0.09 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $2.70 / $13.50 每 1M | $2.70 / $0.27 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $0.16 / $0.96 每 1M | $0.20 / $0.02 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $1.60 / $9.60 每 1M | $2.00 / $0.16 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $3.20 / $16.00 每 1M | $4.00 / $0.32 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 500K | $1.20 / $3.60 每 1M | $0.30 / $0.30 每 1M | 輸入: 文字輸入: 圖像 輸出: 文字 | ||
| 1M | $1.80 / $9.00 每 1M | $2.25 / $0.18 每 1M | 輸入: 文字輸入: 文件 輸出: 文字 | ||
| 1.05M | $0.48 / $0.96 每 1M | $0.10 / $0.10 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1.05M | $1.26 / $3.96 每 1M | $0.23 / $0.23 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| — | $1.00 / $8.00 每 1M | $0.10 / $0.10 每 1M | 輸入: 文字輸入: 圖像輸入: 文件 輸出: 文字 | ||
| 1M | $0.36 / $1.44 每 1M | $0.07 / $0.07 每 1M | 輸入: 文字輸入: 文件 輸出: 文字 | ||
| — | 1.05M | $0.30 / $1.20 每 1M | — / $0.006 每 1M | 輸入: 文字 輸出: 文字 | |
| 262K | $0.85 / $3.60 每 1M | $0.14 / $0.14 每 1M | 輸入: 文字輸入: 文件 輸出: 文字 | ||
| 205K | $0.24 / $0.96 每 1M | $0.30 / $0.02 每 1M | 輸入: 文字輸入: 文件 輸出: 文字 | ||
| 262K | $0.54 / $2.70 每 1M | $0.09 / $0.09 每 1M | 輸入: 文字輸入: 文件 輸出: 文字 | ||
| 262K | $0.10 / $0.97 每 1M | — | 輸入: 文字輸入: 圖像 輸出: 文字 | ||
| 262K | $0.10 / $0.97 每 1M | — | 輸入: 文字輸入: 圖像 輸出: 文字 | ||
| 262K | $0.02 / $0.18 每 1M | — | 輸入: 文字輸入: 圖像 輸出: 文字 |
GPT-6 Sol API
Use GPT-6 Sol when an application needs to reason across files, tool results and several stages of a task. Start with a representative coding or agent workflow, then compare its completed-task rate, latency and total cost with other available models.
Reasoning You Can Tune
Choose none, low, medium, high, xhigh or max according to the job. Start with the default medium setting, then test lower effort for routine work and higher effort for harder tasks.
Coding and Agent Workflows
Use GPT-6 Sol for code changes, debugging and plans that require several steps. Keep tool execution and verification in your application; the model proposes actions, while your workflow determines what runs.
Text and Image Input
OpenAI documents text and image input with text output. Use image understanding for screenshots or diagrams where supported by the selected GPTProto route; this model does not directly output images, audio or video.
What Is the GPT-6 Sol API?
GPT-6 Sol is OpenAI's model for demanding coding and agentic work. Its published model ID is gpt-6-sol. It accepts text and image input, returns text, and supports streaming and structured outputs upstream. It is a hosted proprietary model, not an open-source download.
Sol sits between Luna and the flagship Astra tier. Choose it when coding or multi-step work needs more reasoning than a routine, high-volume task. Test the same work and acceptance criteria across tiers; compare completed tasks, latency and total tokens, not just list prices.
| Published specification | GPT-6 Sol |
|---|---|
| OpenAI model ID | gpt-6-sol |
| Context window | 1,050,000 tokens |
| Maximum output | 128,000 tokens |
| Input / output | Text and image / text |
| Reasoning effort | none, low, medium (default), high, xhigh, max |
| Streaming / structured outputs | Supported upstream |
| Fine-tuning | Not supported |
The context window is a capacity limit, not a reason to send every repository file on each call. Retrieve relevant files and reserve output budget for the patch. For long tasks, log token usage, tool calls and whether the result passed your tests; retries change the true cost.
GPT-6 Sol API Applications
Repository-scale coding. Provide files and interfaces relevant to a change, then check the patch with your tests. This suits cross-file edits, reviews and debugging.
Agentic work. Your application handles search, file access and tool execution. OpenAI recommends Responses for tool calling; Chat Completions supports function calling only with reasoning_effort: "none". Check the route and parameters GPTProto exposes before migrating an agent.
Long documents and visual references. Analyze specifications or screenshots alongside written requirements. Use a separate model when the workflow needs image output.
GPT-6 Sol vs GPT-6 Astra and GPT-6 Luna
All three models publish 1,050,000-token context and 128,000-token maximum output. These are OpenAI upstream standard rates, not GPTProto prices; use the live widget for GPTProto billing.
| Model | Start here when… | OpenAI list input / output per 1M tokens | Practical check |
|---|---|---|---|
| GPT-6 Luna | The workload is focused, repetitive and sensitive to cost or response time. | $0.10 / $0.50 | Measure whether it completes enough tasks without escalation. |
| GPT-6 Sol | Coding, reasoning or multi-step agent tasks need more headroom. | $2 / $10 | Compare accepted results and tool loops against Luna and Astra. |
| GPT-6 Astra | The hardest cross-tool, research or computer-use tasks justify the flagship tier. | $10 / $50 | Check whether its higher success rate reduces total cost per completed task. |
OpenAI publishes separate cached-input, cache-write and long-context rates. Above 272,000 input tokens, higher long-context rates apply to the entire request upstream. Check GPTProto's live billing rules before budgeting a long agent run.
Before Moving an Existing Agent to Sol
Test your prompt and tool schema with the model ID on this page. Confirm that the selected route supports the needed combination of reasoning and tools. Compare low and medium effort on representative tasks, keep a fallback in your application, and track cost per completed task.
FAQ
How much does the GPT-6 Sol API cost on GPTProto?
Use this page's live Pricing panel for GPTProto's input, output and cache rates. OpenAI lists $2 input and $10 output per million tokens for standard shorter-context requests; those upstream rates are not your GPTProto bill.
How do I get a GPT-6 Sol API key and test it?
Create a key in your GPTProto dashboard, then select Try this model to see the model ID and Quick Start. The same GPTProto key works with other supported models. Keep it on the server side.
Can I use GPT-6 Sol for coding and long-running agents?
Yes. OpenAI identifies coding and agents as Sol use cases. Your application manages permissions, tools and tests. Large prompts and repeated tool loops affect cost, so trial a representative task first.
Does GPT-6 Sol support function calling with reasoning enabled?
OpenAI supports function calling via Responses. Its Chat Completions route allows it only with reasoning_effort: "none". Verify GPTProto's endpoint and parameters before depending on reasoning-enabled tool calls.
Is GPT-6 Sol open source, and can it generate images?
No. It is a hosted model. It accepts image input and returns text, but does not generate images. Choose a dedicated model in the OpenAI model collection for image output.
GPT-6 Sol vs Grok 4.7, Opus 5.5 or Gemini 4: which is better?
Compare available versions with identical prompts, tools and acceptance tests. Record completion rate, latency and total cost. Cross-vendor claims need repeatable results; the table above compares only the documented GPT-6 family.
相關文章
與本模型相關的指南、對比與更新。
所有文章
6 Best LLM API Providers in 2026: Multi-Model Platforms Compared
Compare the six best LLM API providers in 2026 by model access, pricing structure, speed, compatibility, governance, and developer experience.
What Is GPT-6 Sol? Release Status, Pricing, Tests, and What We Know
What is GPT-6 Sol? Check its release status, reported features, pricing, coding tests, and how it compares with GPT-6 Astra and GPT-5.6 Sol.
5 Best APIs for Tech Startups in 2026: A Lean MVP Stack
Compare GPTProto, Supabase, Stripe, Clerk, and Resend to build a lean, API-first AI SaaS MVP with lower upfront infrastructure costs.
5 Best Affordable AI Video APIs in 2026: Pricing, Ecommerce, and Short Drama
Compare 5 affordable AI video APIs for ecommerce and AI short drama. See current pricing, clip costs, audio fees, and the best model for each job.