GPT-5.6 Terra vs GPT-6 Astra
什么时候选哪个
gpt-5.6-terra 和 gpt-6-astra 的上下文窗口都是 1050000，最大输出都是 128000，都支持文本和图像输入，思考模式都可以关闭。gpt-5.6-terra 更便宜：输入 $2、输出 $12、缓存输入 $0.2，而 gpt-6-astra 为 $10、$50 和 $1，因此输入和缓存输入便宜约 5 倍，输出便宜约 4.2 倍。想控制成本就选 gpt-5.6-terra；如果需要 2026-04 的知识截止时间（gpt-5.6-terra 为 2026-02），就选 gpt-6-astra。
Benchmark 成绩
20 项两边都有成绩。
供应商公布： Alibaba (Qwen) Anthropic ByteDance DeepSeek Google MiniMax Moonshot OpenAI Tencent Z.ai
价格
| GPT-5.6 Terra | GPT-6 Astra | Δ | |
|---|---|---|---|
| 输入 / 1M token | $2 | $10 | 0.2× |
| 输出 / 1M token | $12 | $50 | 0.24× |
| 缓存读取 / 1M token | $0.2 | $1 | 0.2× |
| 缓存写入 | 不单独收费 | 不单独收费 | - |
价格取自构建时的实时目录，最新价格见各模型页面。
两个模型所处的位置：全部 79 个按同一单位计费的聊天模型的每 1M token 输入价分布（对数刻度）
能力
| GPT-5.6 Terra | GPT-6 Astra | |
|---|---|---|
| 工具调用 | 是 | 是 |
| 思考控制 | 可配置 | 可配置 |
| 结构化输出 | 是 | 是 |
| 提示词缓存 | 隐式（自动） | 隐式（自动） |
| 缓存有效期 | 5-10m, up to 1h | 5-10m, up to 1h |
| 最小缓存前缀 | 1024 个 token | 1024 个 token |
规格
| GPT-5.6 Terra | GPT-6 Astra | |
|---|---|---|
| 输入模态 | 文本 图像 | 文本 图像 |
| 输出模态 | 文本 | 文本 |
| 发布日期 | 2026-07-09 | 2026-09-03 |
| 知识截止日期 | 2026-02 | 2026-04 |
| 上下文窗口 | 1.1M | 1.1M |
| 最大输出 | 128K | 128K |
| 思考参数 | reasoning.effort | reasoning.effort |
| 可选值 | reasoning.effort
| reasoning.effort
|
| 默认值 | medium | medium |
规格照录自各供应商的文档；供应商没有公布的项目，对应的行直接省略，不做推断。 完整来源： GPT-5.6 Terra · GPT-6 Astra
同一条提示词，两个模型，经网关实测
GPT-5.6 Terra 通过 · 3 sentences
Enabling reasoning makes the model generate additional internal analysis tokens and perform more computation before producing the final answer, increasing usage per request. Those extra tokens are billed like other model output or are reflected in higher-priced reasoning-model rates, so even unchanged user traffic can cost substantially more. The bill doubled because average cost per request rose, likely from longer reasoning traces, more complex model routing, or both.
输出 120 tok (+31 思考) 延迟 2.4 s
GPT-6 Astra 通过 · 3 sentences
Enabling reasoning makes the model do more work per request, often generating internal reasoning tokens that are billed even though they do not appear in the final answer. If traffic, model choice, and pricing stayed the same, those extra tokens likely explain why our bill doubled—not longer answers or more users. We should verify the usage breakdown, then reserve higher reasoning levels for tasks where improved results justify the added cost.
输出 283 tok (+196 思考) 延迟 17.2 s
指令遵循（恰好三句，数得出来）、受众适配（对 CFO 说话的口吻），以及下方 token 计数暴露出的隐藏思考计费差额。
GPT-5.6 Terra 通过 · 8/8 cases
```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 best = max(best, cur) return best ``` The bug is that `cur` and `best` were counting adjacent *matches* rather than the number of items in a run: a run of one item has zero matching adjacent pairs, and a run of length `n` has only `n - 1` such pairs. Initializing both counters to `1` for a nonempty list and resetting `cur` to `1` when the value changes makes them represent actual run lengths. Updating `best` on every iteration also ensures a run that continues through the final element is included.
输出 196 tok 延迟 2.6 s
GPT-6 Astra 通过 · 8/8 cases
```python def longest_run(xs): if not xs: return 0 best = cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 best = max(best, cur) return best ``` The original code counts equal adjacent pairs rather than items, so every nonempty run is undercounted by one: a single-element input returns `0`, and a run of three equal items returns `2`. Starting and resetting `cur` to `1` counts the first item of each run, while the empty-input check preserves a result of `0`. Updating `best` after each iteration includes runs ending at the last index without needing a final flush; the original also updated during trailing runs, but undercounted them.
输出 321 tok (+131 思考) 延迟 13.5 s
修复是否真的正确（能运行）、解释的信息密度，以及在一个范围明确的任务上的 token 效率。
GPT-5.6 Terra 通过 · 5/5 fields, guidance "withheld pending the Q4 close"
{"company":"Meridian Robotics","quarter":"September quarter","revenue_usd_m":412,"growth_pct":null,"guidance":"withheld pending the Q4 close"}
输出 149 tok (+103 思考) 延迟 2.3 s
GPT-6 Astra 通过 · 5/5 fields, guidance "Formal guidance is withheld pending the "
{ "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": null, "growth_pct": "eleven and change", "guidance": "Formal guidance is withheld pending the Q4 close." }
输出 329 tok (+267 思考) 延迟 16.9 s
是否严守 schema（不臆造字段）、能否顶住幻觉压力（原文明说暂不给出 guidance），以及结构化输出路径的差异。
GPT-5.6 Terra 通过 · 120 words, 0 banned, 1 question
Today, our API gateway introduces cross-provider prompt caching, reducing repeated token processing while preserving your existing model routing. Teams can store reusable system prompts, policy blocks, and long context prefixes once, then serve them to supported providers with consistent cache keys and observability. The feature tracks hit rates, latency savings, and estimated cost reductions in the dashboard you already use for requests. Why pay to recompute identical context on every call? Configure cache policies by route, tenant, model, or TTL, and fall back automatically when a provider lacks compatible caching. Built-in controls help protect sensitive data through encryption, regional settings, and explicit expiration. Start with a single endpoint, compare results across providers, and scale prompt reuse without rewriting application logic.
输出 631 tok (+475 思考) 延迟 6.0 s
GPT-6 Astra 通过 · 120 words, 0 banned, 1 question
Introducing Cross Provider Prompt Cache, a new API gateway feature that reuses eligible prompt content across supported AI providers. Why pay to process the same context every time? Configure caching once at the gateway, then route requests between models while keeping shared instructions, reference material, and conversation prefixes ready for reuse. Caching controls let teams set expiration windows, isolate tenants, and exclude sensitive content. Cache analytics show hit rates, estimated savings, and latency trends, helping developers tune performance with confidence. Existing routing policies continue to work, so adoption fits your current architecture. Start with a single application, measure the results, and expand as needed. Available today in the dashboard and API, with documentation and examples to guide your first deployment.
输出 665 tok (+516 思考) 延迟 19.3 s
是否守住约束（字数预算、禁用词表、只能有一个问句）、文风特征，以及长度控制。
改一行代码就能在两个模型之间切换
下方每个标签页里都有两个模型 ID，高亮的那两行是唯一要改的地方。端点不变，API key 不变，请求结构也不变。
from openai import OpenAI
client = OpenAI(
base_url="https://synthorai.io/v1",
api_key="sk-syn-...",
)
resp = client.chat.completions.create(
model="gpt-5.6-terra",
# model="gpt-6-astra", # 取消注释此行，注释上一行
messages=[{"role": "user", "content": "Summarize this diff"}],
reasoning_effort="medium",
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://synthorai.io/v1",
apiKey: "sk-syn-...",
});
const resp = await client.chat.completions.create({
model: "gpt-5.6-terra",
// model: "gpt-6-astra", // 取消注释此行，注释上一行
messages: [{ role: "user", content: "Summarize this diff" }],
reasoning_effort: "medium",
});
console.log(resp.choices[0].message.content);curl https://synthorai.io/v1/chat/completions \
-H "Authorization: Bearer sk-syn-..." \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.6-terra",
# "model": "gpt-6-astra", # 取消注释此行，注释上一行
"messages": [{"role": "user", "content": "Hello"}],
"reasoning_effort": "medium"
}'package main
import (
"context"
"fmt"
"github.com/openai/openai-go/v3"
"github.com/openai/openai-go/v3/option"
)
func main() {
client := openai.NewClient(
option.WithBaseURL("https://synthorai.io/v1"),
option.WithAPIKey("sk-syn-..."),
)
resp, _ := client.Chat.Completions.New(context.TODO(), openai.ChatCompletionNewParams{
Model: "gpt-5.6-terra",
// Model: "gpt-6-astra", // 取消注释此行，注释上一行
Messages: []openai.ChatCompletionMessageParamUnion{
openai.UserMessage("Summarize this diff"),
},
ReasoningEffort: openai.ReasoningEffortMedium,
})
fmt.Println(resp.Choices[0].Message.Content)
}import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.*;
import com.openai.models.ReasoningEffort;
OpenAIClient client = OpenAIOkHttpClient.builder()
.baseUrl("https://synthorai.io/v1")
.apiKey("sk-syn-...")
.build();
ChatCompletion resp = client.chat().completions().create(
ChatCompletionCreateParams.builder()
.model("gpt-5.6-terra")
// .model("gpt-6-astra") // 取消注释此行，注释上一行
.addUserMessage("Summarize this diff")
.reasoningEffort(ReasoningEffort.MEDIUM)
.build());
System.out.println(resp.choices().get(0).message().content().orElse(""));常见问题
GPT-5.6 Terra 和 GPT-6 Astra 哪个更便宜？
按「输入 / 1M token」算，GPT-5.6 Terra 更便宜（$2 对 $10，相差 5.0×）。其他计费项的结论可能相反，完整价格见上方表格，实际成本取决于你的用量构成。
不用分别集成两次，就能对 GPT-5.6 Terra 和 GPT-6 Astra 做 A/B 测试吗？
可以。两个模型走同一个 OpenAI 兼容端点，用同一个 API key，切换时只要改一行里的模型名，所以可以给两个模型各分一部分流量，直接对比账单。
GPT-5.6 Terra 和 GPT-6 Astra 支持提示词缓存吗？
支持。两个模型的缓存读取价都低于各自的输入价，所以前缀能反复命中缓存的负载，实际成本会比按官网价估算的低。具体的缓存读取价见上方价格表。