GPT-5
OpenAI · released 2026-01-15
Quality (AA Index)91
Input price$1.25
Output price$10.00
Context400K
Throughput120 tok/s
P50 latency0.7s
GPT-5 vs Claude 4.6 Sonnet
Side-by-side pricing, capabilities, real-world cost across common scenarios, and our editorial pick.
添加模型最多可对比 4 个模型
全部价格按美元 / 百万 token 计
Claude 4.6 Sonnet
Anthropic · released 2026-04-08
Quality (AA Index)89
Input price$3.00
Output price$15.00
Context200K
Throughput85 tok/s
P50 latency1.1s
链接通过我们的 OpenRouter 推荐链接在新标签页打开。 查看推广披露
规格对比
绿色列 = 该指标获胜方| Metric | GPT-5 | Claude 4.6 Sonnet | Verdict |
|---|---|---|---|
Input price /百万 tokens | $1.25 | $3.00 | GPT-5 −58% |
Output price /百万 tokens | $10.00 | $15.00 | GPT-5 −33% |
Context window max input length | 400K | 200K | GPT-5 +2.0× |
AA Quality AA Intelligence Index (0–100) | 91 | 89 | GPT-5 +2pt |
Arena Elo LMArena human-pref Elo (800–2000) | — | — | Tied |
Throughput tokens per second | 120 | 85 | GPT-5 +41% |
P50 latency first token | 0.7s | 1.1s | GPT-5 −36% |
Vision multimodal | Tied | ||
Function calling tool use | Tied | ||
Reasoning mode chain-of-thought | Tied |
常见场景月度费用
默认用量假设| Scenario | GPT-5 | Claude 4.6 Sonnet |
|---|---|---|
customer support 1000K req · 600/180 tok | $2.5K | $4.5K |
chat with docs 300K req · 4000/300 tok | $2.4K | $5.0K |
code generation 500K req · 2000/500 tok | $3.8K | $6.8K |
voice assistant 600K req · 800/200 tok | $1.8K | $3.2K |
Our pick
For most workloads, choose GPT-5.
- 58% cheaper input price, which compounds at scale
- 2.0× the context window — better for long documents and agents
- 41% faster throughput — matters for streaming UX and voice agents
Best across-the-board quality but the most expensive token rate in the table; pick GPT-5 when reasoning depth or function-calling reliability outranks per-million-token spend.
Choose Claude 4.6 Sonnet instead if: Strongest non-OpenAI option for nuanced writing and code review; pricier than DeepSeek but more reliable refusal patterns for regulated industries.
阅读完整深度分析
为何选择？通过智能路由同时使用两者
第二阶段 · 带故障转移链的网关Set GPT-5 as primary, Claude 4.6 Sonnet as fallback. One key, one bill, automatic failover when GPT-5 errors.
第二阶段预览 · 网关尚未上线该接口目前不存在。网关计划在第二阶段上线——下面只是规划中的接口形态预览，不是可用的 API。上线时会通过 newsletter 通知订阅者。
查看计划中的 API 调用形态
$ curl https://api.aipricly.com/v1/chat/completions \
-H "Authorization: Bearer $AIPC_KEY" \
-d '{
"routing": {
"primary": "openai/gpt-5",
"fallback": ["anthropic/claude-4-6-sonnet"]
},
"messages": [{"role": "user", "content": "..."}]
}'