← Every GPUNVIDIA · 141 GB VRAM · 4800 GB/s

What can the H200 run?

With 141 GB, 10 of 12 popular open models fit. The best pick is gpt-oss 120B at about 580 tok/s, faster than you can read.

ollama run gpt-oss:120b

First time? Install Ollama, then paste that into a terminal. Is this your machine? Check it on the homepage.

141 GB VRAM

Every model the H200 can run

  • Qwen3 4BQuick everyday chat and summaries
    ~1060 tok/s
    Runs greatqwen3:4b
  • Qwen3 8BA solid all-rounder for chat and writing
    ~510 tok/s
    Runs greatqwen3:8b
  • Gemma 3 12BReads images as well as text
    ~330 tok/s
    Runs greatgemma3:12b
  • Qwen3 14BNoticeably smarter at reasoning and code
    ~280 tok/s
    Runs greatqwen3:14b
  • gpt-oss 20BOpenAI's open model. Fast, strong reasoning
    ~840 tok/s
    Runs greatgpt-oss:20b
  • Gemma 3 27BGoogle's best open model for its size
    ~160 tok/s
    Runs greatgemma3:27b
  • Qwen3 30B-A3B30B smarts at small-model speed
    ~840 tok/s
    Runs greatqwen3:30b
  • Qwen3 32BStrong coding and long answers
    ~130 tok/s
    Runs greatqwen3:32b
  • Llama 3.3 70BClose to the big cloud models
    ~61 tok/s
    Runs greatllama3.3:70b
  • gpt-oss 120BOpenAI's largest open model
    ~580 tok/s
    Runs greatgpt-oss:120b
  • Qwen3 235B-A22BAlibaba's flagship open model, for multi-GPU rigs
    needs 144 GB
    Too bigqwen3:235b
  • DeepSeek R1 671BThe full DeepSeek R1, for the biggest rigs
    needs 406 GB
    Too bigdeepseek-r1:671b
Common questions

Can the H200 run it?

Can the H200 run Llama 3.3 70B?

Yes. Llama 3.3 70B needs about 45 GB and the H200 has 141 GB of VRAM. Expect about 61 tok/s, faster than you can read.

Can the H200 run gpt-oss 120B?

Yes. gpt-oss 120B needs about 67 GB and the H200 has 141 GB of VRAM. Expect about 580 tok/s, faster than you can read.

Can the H200 run Qwen3 235B-A22B?

No. Qwen3 235B-A22B needs about 144 GB and the H200 has at most 141 GB of VRAM. Ollama can still start it by putting part of the model in system RAM, but it runs several times slower. The most capable model it runs well is gpt-oss 120B, at about 580 tok/s.

Measured, not estimated

Real runs on the H200

Loading measured runs…

How we estimate

Generating a token reads the model's active weights from memory once, so speed is roughly memory bandwidth (4800 GB/s) divided by the active weight size, with efficiency calibrated on our own rack. Sizes are Ollama's default 4-bit downloads plus about 1.2 GB for context. Real results are usually within about a third. Measured runs are on the leaderboard.

Want something bigger?

The Dell DSS 8440 + 10× Tesla P100 runs Qwen3 235B-A22B at about 19 tok/s, from $5,100 used.

See the builds

Compare

Similar cards

Every GPU we track

Head to head

The H200 against its rivals

Every comparison