Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. Models
  2. Text
  3. Google

Gemma 4 31B Thinking

Google's Gemma 4 31B instruction-tuned model with thinking explicitly enabled, exposing reasoning traces for complex multimodal and coding workflows. Requests containing video cost $0.14 per million input tokens and $0.40 per million output tokens.

  • Reasoning
  • Vision
  • Video Input
  • Structured Output

Added Apr 2, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.10
Output
$0.34
Cache read
$0.10
Compare provider prices

Specifications

Context window
262.1K
Max output
131.1K
Parameters
31B

Benchmarks

Sourced from Vectara.

Hallucination Rate

7.4%

Hallucination Rate: 7.4
Better than 67% of models compared

Factual Consistency

92.6%

Factual Consistency: 92.6
Better than 67% of models compared

Answer Rate

100.0%

Answer Rate: 100.0
Better than 91% of models compared

Avg Summary Length

Average generated summary length

75.8

Last updated 2026-10-22 · Matched as google/gemma-4-31b-it

Vectara Leaderboard

Leaderboard results

These leaderboards tested Gemma 4 31B, another version of this model.

BenchmarkScoreRank on NanoGPT
Epoch AI Benchmarking Hub

Epoch AI data, CC BY 4.0 · Updated Nov 13, 2026

Epoch Capabilities Index

144.3

#84 of 115

GPQA Diamond

75.8%

#87 of 119

SimpleQA Verified

10.4%

#80 of 82

OTIS Mock AIME 2024–2025

73.3%

#85 of 121

EQ-Bench

EQ-Bench data, MIT · Updated Nov 13, 2026

EQ-Bench 4

1110

#25 of 31

Creative Writing v3

1369

#69 of 96

Longform Writing

56.5

#62 of 97

EQ-Bench 3

1464

#22 of 49

UGI Leaderboard

UGI Leaderboard data, Apache-2.0 · Updated Oct 16, 2026

UGI Score

23.1

#121 of 149

UGI Willingness

2.5

#94 of 149

UGI Writing

43.1

#72 of 149

Filtered and ranked to models on NanoGPT.

Providers

Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)

Auto
Auto (ZDR)
$0.10$0.34$0.10N/AN/A
Auto
Auto (Anonymized)
$0.10$0.34$0.10N/AN/A
Sail Research (Standard)
Sail Research (Standard)
FP4
GLB
$0.0735$0.42$0.0525N/AN/A
Creekside Compute
Creekside Compute
NVFP4
GLB
$0.0954$0.278$0.0112N/AN/A
Runware
Runware
10% off
UNK
$0.107$0.0964$0.312$0.281$0.0126$0.0113N/AN/A
CoreWeave
CoreWeave
FP4
US
$0.105$0.357$0.105N/AN/A
Hyperfusion
Hyperfusion
BF16
AE
$0.136$0.399N/AN/AN/A
Ollama
Ollama
UNK
$0.147$0.42$0.0525N/AN/A
Sail Research (ASAP)
Sail Research (ASAP)
FP4
GLB
$0.147$0.42$0.0735N/AN/A
Crusoe
Crusoe
BF16
US
$0.147$0.42$0.147N/AN/A
DekaLLM
DekaLLM
ID
$0.147$0.42$0.0525N/AN/A
Friendli
Friendli
US
$0.147$0.42N/AN/AN/A
Novita
Novita
BF16
US
$0.147$0.42N/AN/AN/A
NeuralWatt
NeuralWatt
FP4
US
$0.151$0.441$0.0151N/AN/A
NextBit
NextBit
FP4
EU
$0.157$0.42$0.063N/AN/A
Parasail
Parasail
FP8
US
$0.157$0.42$0.063N/AN/A
PrimeIntellect
PrimeIntellect
FP4
UNK
$0.21$0.352$0.105N/AN/A
SCX (Global)
SCX (Global)
BF16
GLB
$0.315$0.956N/AN/AN/A
Uomi
Uomi
FP8
GLB
$0.344$0.60N/AN/AN/A
Modal
Modal
US
$0.357$0.105$0.105N/AN/A
SambaNova
SambaNova
US
$0.399$1.21N/AN/AN/A
Tinfoil
Tinfoil
TEE verified
US
$0.42$1.05N/AN/AN/A
SiliconFlow
SiliconFlow
FP8
SG
$0.787$1.05$0.263N/AN/A
Privacy:ZDR, prompts not retainedNo training on promptsRetention and training unclear
Batch API
ProviderInputOutputCached input
Auto
Autogoogle/gemma-4-31b-it:thinking
$0.10/M$0.34/M$0.10/M

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Related text models

Compare Gemma 4 31B Thinking with similar models from the same provider or model family.

Gemma 4 31B

google/gemma-4-31b-it

Google's Gemma 4 31B instruction-tuned model for heavier reasoning, coding, agentic workflows, and long-context multimodal understanding. This route keeps tokenizer thinking disabled for faster direct answers. Requests containing video cost $0.14 per million input tokens and $0.40 per million output tokens.

Gemma 4 26B A4B Cybersecurity

google/gemma-4-26b-a4b-it-cybersecurity

Gemma 4 26B A4B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.

Gemma 4 26B A4B

google/gemma-4-26b-a4b-it

Google's Gemma 4 26B A4B instruction-tuned model built for scalable reasoning, coding, long-context, and multimodal workflows. This route is tuned for faster direct answers while preserving multimodal and structured output support. Requests containing video cost $0.10 per million input tokens and $0.40 per million output tokens.

Gemma 4 26B A4B Thinking

google/gemma-4-26b-a4b-it:thinking

Google's Gemma 4 26B A4B instruction-tuned model with structured reasoning for more deliberate coding, multimodal analysis, and long-context problem solving. Requests containing video cost $0.10 per million input tokens and $0.40 per million output tokens.

DiffusionGemma

google/diffusiongemma

DiffusionGemma is a high-speed diffusion-based version of Gemma 4 26B A4B. It supports optional reasoning and a 262,144-token context window. Temperature, min-p, seed, minimum output tokens, logit bias, bad words, and allowed token IDs are unavailable; omit these parameters to use native diffusion sampling.

Gemma 4 31B Scarlet

gemma-4-31b-it-scarlet

Scarlet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.