Gemma 4 31B Instruct is a 30.7B dense model with a long context window, multilingual performance, function calling, and configurable reasoning. Running inside a TEE (Trusted Execution Environment), with provider attestation support.
Added May 26, 2026
Model weightsContext Window
262.1K
Max Output
262.1K
Input Price (Auto)
$0.15/1M
Output Price (Auto)
$0.46/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
19.0
Coding Index
43.4
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
4.9%
Better than 28% of models compared
Harvey LAB-AA
Legal agentic work criterion pass rate
47.2%
Better than 4% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
362 Elo
Better than 18% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
606 Elo
Better than 33% of models compared
Document reasoning
AA-LCR v1.1
Long context reasoning with updated grading
69.7%
Better than 63% of models compared
Reasoning
HLE
Humanity's Last Exam
23.6%
Better than 73% of models compared
IFBench
Instruction-following benchmark
75.6%
Better than 93% of models compared
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
0.0%
Better than 16% of models compared
SciCode
Python programming for scientific computing
45.5%
Better than 42% of models compared
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
85.7%
Better than 79% of models compared
Terminal-Bench Hard (legacy)
Agentic coding and terminal use
36.4%
Better than 83% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
59.9%
Better than 56% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
69.7%
Better than 63% of models compared
Last updated Sep 24, 2026
Artificial AnalysisProviders
Provider information for this model’s automatic routing. These routes cannot be selected individually.
Loading provider options…
Related text models
Compare Gemma 4 31B IT TEE with similar models from the same provider or model family.
Gemma 4 31B Thinking TEE
TEE/gemma4-31b:thinkingGoogle's Gemma 4 31B instruction-tuned model in a Trusted Execution Environment (TEE) with thinking explicitly enabled and provider attestation support.
Gemma 4 31B TEE
TEE/gemma4-31bGoogle's Gemma 4 31B instruction-tuned model in a Trusted Execution Environment (TEE) with attestation support. This route keeps reasoning disabled for faster direct answers while preserving structured output support.
Gemma 4 26B A4B Uncensored TEE
TEE/gemma-4-26b-a4b-uncensoredGemma 4 26B A4B Uncensored Heretic is an MoE model tuned to reduce refusal behavior while preserving coding, reasoning, and function-calling strengths. Running inside a TEE (Trusted Execution Environment), with provider attestation support. Image input is unavailable on this deployment.
Gemma 4 31B Split-Untied
google/gemma4-31b-splituntiedBlazed-Forge's Split-Untied is a text-only Gemma 4 31B community finetune with an untied BF16 output head, built for creative writing, roleplay, expressive dialogue, and tool use.
Gemma 4 31B DarkIdol
gemma-4-31b-it-darkidolDarkIdol is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.
Gemma 4 31B Fabled
gemma-4-31b-it-fabledFabled is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.