Head to head

Gemma 4 31B vs gpt-oss-120b

Gemma 4 31B scores higher on 3 of the 4 benchmarks both models publish. The widest gap is on Humanity's Last Exam (with tools), where Gemma 4 31B leads by 7.5 points. Both have open weights; Gemma 4 31B is far lighter to host, needing about 21 GB at 4-bit against 81 GB for gpt-oss-120b.

Gemma 4 31BGoogle · OpenLLM Buddy catalog79.2compositegpt-oss-120bOpenAI · Open weights57.3composite

Shared benchmarks

Raw published scores; the higher score on each row is highlighted. LMArena Elo is a human-preference vote and is not counted as a win.

Should you pick Gemma 4 31B or gpt-oss-120b?

Pick Gemma 4 31B if…

  • You care most about GPQA Diamond, Humanity's Last Exam, Humanity's Last Exam (with tools), where it scores at least 2 points higher.
  • You want the lighter model to host: about 21 GB at 4-bit against 81 GB.
  • You want to run it locally: at 4-bit it fits one 24–32 GB consumer GPU.
  • You want it running in minutes on a dedicated GPU through OpenLLM Buddy, in an uncensored build.

Pick gpt-oss-120b if…

  • You care most about AIME, where it scores at least 2 points higher.
  • You need fast responses: about 212 tokens per second through APIs, against 35.
  • Cost matters: $0.150 per million output tokens against $0.340.

Hardware and cost to run

Gemma 4 31B

At 4-bit it needs about 21 GB, which fits a single RTX 4090, so it can run on a home workstation with llama.cpp, Ollama, or vLLM. Full BF16 precision needs about 71 GB (1× H100 (80 GB)).

On OpenLLM Buddy: gemma-4-31B-it-uncensored runs on RTX 4090 24GB, with an OpenAI-compatible API and no per-token fees. Prices start at $13 USD for 11 hours.

Deploy Gemma 4 31B
Full GPU memory table for Gemma 4 31B

gpt-oss-120b

Even at 4-bit it needs about 81 GB (1× H200 (141 GB)), well beyond a home workstation, so most people run it on rented multi-GPU servers. Full BF16 precision needs about 269 GB (4× H100 (320 GB)).

Full GPU memory table for gpt-oss-120b

Category profile

Category index uses every test each model reports, not only the shared ones.

ReasoningCodingAgenticKnowledge
Gemma 4 31Bgpt-oss-120b
Reasoning69.868.2
Coding85.676.5
Agentic68.927.1
Knowledge92.5—

Specs side by side

Gemma 4 31Bgpt-oss-120b
LabGoogleOpenAI
WeightsOpenOpen
Released2026-04-022025-08-05
Parameters30.7B dense117B total, 5.1B active (MoE)
Context256K128K
LicenseApache-2.0Apache 2.0
4-bit memory21 GB · RTX 4090 (24 GB)81 GB · 1× H200 (141 GB)
API price$0.100 in / $0.340 out per 1M tokens$0.030 in / $0.150 out per 1M tokens
API speed35 tok/s212 tok/s
One-GPU speed (est.)≈ 25–34 tok/s on RTX 4090 (4-bit)—

Related comparisons

Sources

ModelTestScoreSettingSource
Gemma 4 31BGPQA Diamond84.3—Gemma 4 31B IT model card (Hugging Face)
gpt-oss-120bGPQA Diamond80.1high effort, no toolsgpt-oss model card paper (arXiv 2508.10925), Table 3
Gemma 4 31BHumanity's Last Exam19.5no toolsGemma 4 31B IT model card (Hugging Face)
gpt-oss-120bHumanity's Last Exam14.9high effort, no toolsgpt-oss model card paper (arXiv 2508.10925), Table 3
Gemma 4 31BHumanity's Last Exam (with tools)26.5with searchGemma 4 31B IT model card (Hugging Face)
gpt-oss-120bHumanity's Last Exam (with tools)19.0high effort, with toolsgpt-oss model card paper (arXiv 2508.10925), Table 3
Gemma 4 31BAIME89.2AIME 2026, no toolsGemma 4 31B IT model card (Hugging Face)
gpt-oss-120bAIME95.8AIME 2024, high effort, no toolsgpt-oss model card paper (arXiv 2508.10925), Table 3

Scores are published results from model cards, lab reports, and public leaderboards, each linked to its source. Settings such as thinking mode or test version differ between sources and are shown next to each score. OpenLLM Buddy catalog models are uncensored fine-tunes; their scores are those of the official base model. Data collected 2026-10-30. How the index works