# DeepSeek R1

DeepSeek R1 je AI model od DeepSeek s otevřenými váhami, vydaný 2025-01-20. Kontextové okno má velikost 64 000 tokenů.

- Vývojář: DeepSeek
- Id: deepseek-ai/DeepSeek-R1
- Podle zpráv: 2026-10-07T20:58:27.000Z
- Vydáno: 2025-01-20T00:00:00.000Z
- Zjištění: 2026-10-07T19:40:57.518Z
  - [zdrojový odkaz](https://huggingface.co/deepseek-ai/DeepSeek-R1)
- Stav: Vydáno
- Verze vah: 2025-01-20
  - [zdrojový odkaz](https://huggingface.co/api/models/deepseek-ai/DeepSeek-R1/paths-info/56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad)
- Kanonický: https://alexgetman.com/cs/models/deepseek/deepseek-r1/

## Otázky

### Jaké je kontextové okno DeepSeek R1?

Kontextové okno má velikost 64 000 tokenů.

### Jsou váhy modelu DeepSeek R1 veřejně dostupné?

Ano. DeepSeek zveřejňuje váhy modelu.

### Kdy byl vydán DeepSeek R1?

2025-01-20.

## Časová osa

- 2026-11-10T17:59:50.472Z · [GitHub · openai/openai-agents-python · code](https://github.com/openai/openai-agents-python/blob/64fbde5d02b42ae5f9db668eb0184cdb3fc9d775/tests/models/test_openai_chatcompletions_stream.py)
>          object="chat.completion.chunk",
>             choices=[Choice(index=0, delta=delta)],
>         )
>         for delta in [*reasoning_deltas, ChoiceDelta(content="answer")]
>     ]
> 
>     events = await _collect_handler_events(*chunks, model="deepseek-r1")
>     completed_event = next(event for event in events if event.type == "response.completed")
>     reasoning_item = completed_event.response.output[0]
>     assert isinstance(reasoning_item, ResponseReasoningItem)
>     assert reasoning_item.summary[0].text == "summary"
>     assert reasoning_item.content
>     assert [part.text for part in reasoning_item.content] == ["raw"]
>     assert "_chat_completions_reasoning_field" not in cast(Any, reasoning_item).provider_data
> 
>     replayed_messages = Converter.items_to_messages(
>         [item.model_dump() for item in completed_event.response.output],
>         model="deepseek-r1",
>     )
>     assert len
Tato zmínka v kódu nepotvrzuje používaný identifikátor modelu.
- 2026-11-13T18:47:56.425Z · [GitHub search · DeepSeek · pulls · NVIDIA/TensorRT-LLM · @farazkh80](https://github.com/NVIDIA/TensorRT-LLM/pull/20229)
> with all of its KV remote and MLA (DeepSeek-R1 shape) 1.43x, which is why allocation and placement keep pages local. Partitioned attention rejects FlashMLA, beam
- 2026-11-13T18:47:56.425Z · [GitHub search · DeepSeek · pulls · NVIDIA/TensorRT-LLM · @moraxu](https://github.com/NVIDIA/TensorRT-LLM/pull/22139)
- 2026-11-13T18:47:56.425Z · [GitHub search · DeepSeek · pulls · ROCm/aiter · @karverma-amd](https://github.com/ROCm/aiter/pull/6234)
> Tuned configs for DeepSeek-R1 MLA-absorb bmms (B=128). Previously fell back to DEFAULT (untuned): 5.5x slower at decode M~1536 . Lets the BF16 MLA-absorb path run
- 2026-11-13T18:47:56.425Z · [GitHub search · DeepSeek · pulls · anomalyco/models.dev · @models-dev-automation[bot]](https://github.com/anomalyco/models.dev/pull/13568)
> Instruct-Turbo`, `openai/gpt-oss-120b-Turbo`, `google/gemma-4-31B-it-Ultra`, `mattshumer/Reflection-Llama-3.1-70B`, `NovaSky-AI/Sky-T1-32B-Preview`, `deepseek-ai/DeepSeek-R1`, `Qwen/Qwen2-7B-Instruct`, `Qwen/QwQ-32B`, `google/gemma-2-9b-it`, `google/gemma-4-31B-it-turbo`, `Qwen/Qwen3.5-0.8B`, `deepseek-ai/DeepSeek-V3-0324-Turbo`, `openbmb/MiniCPM-Llama3-V-2_5`, `microsoft/phi-4`, `inclusionAI/Ling-3.0-flash-VL`, `Qwen/Qwen3-VL-30B-A3B-Thinking`, `MiniMaxAI/MiniMax-M2.7-Turbo`, `deepseek-ai
- 2026-11-13T18:47:56.425Z · [GitHub search · DeepSeek · pulls · overnightworks/speech · @dependabot[bot]](https://github.com/overnightworks/speech/pull/354)
> XTX, <a href="https://redirect.github.com/vllm-project/vllm/issues/59774">#59774</a>); and gfx1250 enablement for gpt-oss-120b, DeepSeek-R1, DeepSeek-V4-Flash and MiniMax-M3 (<a href="https://redirect.github.com/vllm-project/vllm/issues/58800">#58800</a>).</li>
- 2026-11-13T18:48:53.232Z · [GitHub search · DeepSeek · commits · warejohn4366/war · @warejohn4366](https://github.com/warejohn4366/war/commit/190f9649388346129f29790e242affbec2dfbef0)
> 随着AI发展，企业和开发者希望快速应用AI模型。本文介绍在阿里云PAI Model Gallery上零代码一键部署DeepSeek-V3和DeepSeek-R1模型的方法，包括准备工作、部署步骤、调用方式等。这种方式简化流程、提高效率、降低成本，助力AI能力落地。

## Zdrojové odkazy

- [Arena · leaderboards](https://api-docs.deepseek.com/news/news250120)
- [review](https://api-docs.deepseek.com/news/news250120/)
- [Artificial Analysis · benchmarks](https://artificialanalysis.ai/models)
- [vertex-model-garden](https://console.cloud.google.com/vertex-ai/publishers/deepseek-ai/model-garden/deepseek-r1)
- [AWS Bedrock · lifecycle](https://docs.aws.amazon.com/en_en/bedrock/latest/userguide/model-lifecycle-legacy.html)
- [GitHub search · DeepSeek · pulls · NVIDIA/TensorRT-LLM · @farazkh80](https://github.com/NVIDIA/TensorRT-LLM/pull/20229)
- [GitHub search · DeepSeek · pulls · NVIDIA/TensorRT-LLM · @moraxu](https://github.com/NVIDIA/TensorRT-LLM/pull/22139)
- [GitHub search · DeepSeek · pulls · ROCm/aiter · @karverma-amd](https://github.com/ROCm/aiter/pull/6234)
- [GitHub search · DeepSeek · pulls · anomalyco/models.dev · @models-dev-automation[bot]](https://github.com/anomalyco/models.dev/pull/13568)
- [GitHub · openai/openai-agents-python · code](https://github.com/openai/openai-agents-python/blob/64fbde5d02b42ae5f9db668eb0184cdb3fc9d775/tests/models/test_openai_chatcompletions_stream.py)
- [GitHub search · DeepSeek · pulls · overnightworks/speech · @dependabot[bot]](https://github.com/overnightworks/speech/pull/354)
- [GitHub search · DeepSeek · commits · warejohn4366/war · @warejohn4366](https://github.com/warejohn4366/war/commit/190f9649388346129f29790e242affbec2dfbef0)
- [hf-author-artifacts:deepseek-ai](https://huggingface.co/api/models/deepseek-ai/DeepSeek-R1/paths-info/56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad)
- [huggingface-router](https://huggingface.co/deepseek-ai/DeepSeek-R1)
- [models.dev · catalogue](https://models.dev/#deepseek-r1)
- [OpenCode · discovery](https://opencode.ai/data/deepseek/deepseek-r1)
- [OpenRouter](https://openrouter.ai/deepseek/deepseek-r1)
- [OpenRouter · usage](https://openrouter.ai/rankings)

## Signály

- [DeepSeek R1: Katalogový výpis](https://alexgetman.com/cs/signals/2026/09/11/deepseek-r1/)