64K context · 671B params · MIT
deepseek/deepseek-r1
DeepSeek R1 is here: Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass....
Reachability & verification
Live reachability
Target (DeepSeek): https://api.deepseek.com/v1
Live reachability probe
Measured in real time from the ModelHub gateway egress to this model's API endpoint. Green means reachable right now. Overseas reference latencies are estimates; the live probe is the proof of current reachability.
Pricing across providers
2Prices are shown in each provider's billing currency — ¥ (CNY) for Chinese providers, $ (USD) for overseas gateways.
| Provider | Input /1M | Output /1M | Blended | Ctx | p50 latency | API | Verified |
|---|---|---|---|---|---|---|---|
| DeepSeekdeepseek-r1 | ¥0.7 | ¥2.5 | ¥1.15 | 64K | 920 ms | OAI | 119d ago |
| Together.aideepseek-ai/DeepSeek-R1 | $3 | $7 | $4 | 33K | 350 ms | OAI | 140d ago |
Works with
Any OpenAI-compatible client works — point the base URL at the provider endpoint.
Capabilities
Languages: en, zh
Benchmarks
| Benchmark | Setting | Verification | Score |
|---|---|---|---|
| MMLU | 5-shot | official | 90.3% |
| HumanEval | 0-shot pass@1 | official | 89.1% |
Code samples
from openai import OpenAI
client = OpenAI(
base_url="https://api.deepseek.com/v1",
api_key="YOUR_API_KEY",
)
resp = client.chat.completions.create(
model="deepseek-r1",
messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)curl https://api.deepseek.com/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-r1",
"messages": [{"role": "user", "content": "Hello!"}]
}'import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.deepseek.com/v1",
apiKey: "YOUR_API_KEY",
});
const resp = await client.chat.completions.create({
model: "deepseek-r1",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(resp.choices[0].message.content);Technical specs
Context
64K
Max output
16K
Parameters
671B
Release
2025-01-20
Training cutoff
2024-10-01
License
MIT
Similar models
Frequently asked questions
How much does R1 cost?
The cheapest tracked blended price is ¥1.15 per 1M tokens on DeepSeek. See the pricing matrix above for input/output splits per provider.
Can I use R1 from outside China?
Availability depends on the hosting provider. Use our Path Planner on the DeepSeek page to map a verified official-vs-brokered access path, including payment and KYC constraints.
Is R1 open source?
Yes — R1 is open-weight under the MIT license. You can self-host it or use any listed inference provider.
Is R1 OpenAI-compatible?
Yes — at least one tracked provider exposes an OpenAI-compatible endpoint, so Cursor, Cline, Aider and similar clients work with just a base-URL change.
What is the maximum context window?
R1 supports up to 64K tokens of context with a maximum output of 16K tokens.
Need verified access to R1?
We broker vetted intros to DeepSeek — payment, KYC and endpoint verification handled.