Baltor Get started

Models

DeepSeek-R1-Distill-Llama-70B

By deepseek-ai. 70.6 billion parameters, a context window of 131,072 tokens and the licence mit.

Facts

Released
2025-01-20 the day the repository was first published on Hugging Face Hugging Face, read
Licence
mit the licence the model card declares Hugging Face, read
Open weights
Yes the weights are published in this Hugging Face repository Hugging Face, read
Parameters
70.6 billion counted from the safetensors weight files Hugging Face, read
Active parameters
Unknown
Knowledge cutoff
Unknown
Context window
  • 131,072 tokens the maximum position embeddings in the model configuration model configuration, read
Longest output
Unknown
Tool calling
  • No the chat template in the tokenizer configuration has no place for tool definitions Hugging Face, read
Structured output
Unknown
Reasoning controls
Unknown
Inputs and outputs
text in, text out Hugging Face, read
Good for
No source names a use.

Prices

Each provider's own price in US dollars per million tokens, as an openly licensed source lists it
ProviderInputOutputContextAs ofSource
FastRouter0.030 USD0.140 USD131,072 older than 30 days; check the providermodels.dev, read
Deep Infra0.200 USD0.600 USD131,072LiteLLM, read
Nebius Token Factory0.250 USD0.750 USD128,000LiteLLM, read
nscale0.375 USD0.375 USDUnknownLiteLLM, read
Together AI2.00 USD2.00 USD131,072LiteLLM, read

Run it on your own hardware

Weights are the sizes of the files a source lists, or an estimate from the parameter count where none does. Memory is an estimate: weights plus KV cache plus 512 MiB and 5 percent of the weights for runtime buffers. Check it against your hardware.

QuantizationWeightsMemory at 8,192 tokensMemory at 32,768 tokens
UD-IQ1_S14.8 GiB18.5 GiB26.0 GiB
UD-IQ1_M16.0 GiB19.8 GiB27.3 GiB
UD-IQ2_XXS18.1 GiB22.0 GiB29.5 GiB
UD-IQ2_M22.7 GiB26.8 GiB34.3 GiB
Q2_K24.6 GiB28.8 GiB36.3 GiB
Q2_K_L24.8 GiB29.0 GiB36.5 GiB
UD-Q2_K_XL25.1 GiB29.4 GiB36.9 GiB
UD-IQ3_XXS25.8 GiB30.0 GiB37.5 GiB
Q3_K_S28.8 GiB33.2 GiB40.7 GiB
Q3_K_M31.9 GiB36.5 GiB44.0 GiB
UD-Q3_K_XL32.5 GiB37.1 GiB44.6 GiB
IQ4_XS35.3 GiB40.1 GiB47.6 GiB
IQ4_NL37.3 GiB42.2 GiB49.7 GiB
Q4_037.4 GiB42.2 GiB49.7 GiB
Q4_K_S37.6 GiB42.5 GiB50.0 GiB
Q4_K_M39.6 GiB44.6 GiB52.1 GiB
UD-Q4_K_XL39.7 GiB44.7 GiB52.2 GiB
Q4_141.3 GiB46.3 GiB53.8 GiB
Q5_K_S45.3 GiB50.6 GiB58.1 GiB
Q5_K_M46.5 GiB51.8 GiB59.3 GiB
UD-Q5_K_XL46.5 GiB51.9 GiB59.4 GiB
UD-Q6_K_XL57.0 GiB62.8 GiB70.3 GiB
Q8_069.8 GiB76.3 GiB83.8 GiB
UD-Q8_K_XL75.7 GiB82.4 GiB89.9 GiB
BF16131 GiB141 GiB148 GiB
F16131 GiB141 GiB148 GiB

Start it with a local runtime

  • Ollama
    ollama run hf.co/unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF:Q4_0
  • LM Studio
    lms get unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF
    lms server start
  • llama.cpp server
    llama-server -hf unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF:Q4_0 --jinja
  • vLLM
    vllm serve deepseek-ai/DeepSeek-R1-Distill-Llama-70B
  • SGLang
    sglang serve --model-path deepseek-ai/DeepSeek-R1-Distill-Llama-70B --port 30000

Use it from your harness

Set up for DeepInfra with the model deepseek-ai/DeepSeek-R1-Distill-Llama-70B. Each endpoint page has the same setup for its own address.

OpenCode

Put this in opencode.json in your project folder:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "deepinfra": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "DeepInfra",
      "options": {
        "baseURL": "https://api.deepinfra.com/v1/openai",
        "apiKey": "{env:DEEPINFRA_TOKEN}"
      },
      "models": {
        "deepseek-ai/DeepSeek-R1-Distill-Llama-70B": {
          "name": "DeepSeek-R1-Distill-Llama-70B"
        }
      }
    }
  }
}
  • OpenCode reads any OpenAI-compatible address through the @ai-sdk/openai-compatible package, and an address that speaks the Responses API through @ai-sdk/openai.

From OpenCode documentation, read .

Pi

Put this in ~/.pi/agent/models.json:

{
  "providers": {
    "deepinfra": {
      "baseUrl": "https://api.deepinfra.com/v1/openai",
      "api": "openai-completions",
      "apiKey": "$DEEPINFRA_TOKEN",
      "models": [
        {
          "id": "deepseek-ai/DeepSeek-R1-Distill-Llama-70B"
        }
      ]
    }
  }
}
  • The apiKey field can name an environment variable as $NAME.

From Pi documentation, read .

Codex

Codex speaks only the Responses API, and DeepInfra documents no Responses address. A gateway that offers one can sit in between.

  • Codex speaks the Responses API only: responses is the one supported wire API of a custom provider. Ollama and LM Studio are built in and start with --oss.

From Codex documentation, read .

Claude Code

Put this in ~/.claude/settings.json, or variables in your shell:

export ANTHROPIC_BASE_URL="https://api.deepinfra.com/anthropic"
export ANTHROPIC_AUTH_TOKEN="$DEEPINFRA_TOKEN"
export ANTHROPIC_MODEL="deepseek-ai/DeepSeek-R1-Distill-Llama-70B"
claude
  • Claude Code sends Anthropic Messages requests to ANTHROPIC_BASE_URL. Anthropic says it does not support routing Claude Code to models other than Claude through any gateway, so some features may not work with another model.

From Claude Code documentation, read .

Other ways to reach it

  • FastRouter, model deepseek-ai/deepseek-r1-distill-llama-70b
  • Nebius Token Factory, model deepseek-ai/DeepSeek-R1-Distill-Llama-70B
  • Together AI, model deepseek-ai/DeepSeek-R1-Distill-Llama-70B
  • Ollama, model hf.co/unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF:Q4_0

Published results

These are results other people published. Baltor did not run them and does not rank models by them.

Two more publishers are linked, never copied, because their terms do not allow it:

  • Artificial Analysis publishes its own intelligence, speed and price measurements.
  • OpenRouter lists current prices and uptime for each provider it routes to.