Gemma 4 31B vs Grok-3
Gemma 4 31B and Grok-3 are closely matched at 33.1 and 30.4 on the LLM Stats Score. Gemma 4 31B is 39.3x cheaper per token.
Google · xAI · Updated for 2026
Which is better?
Gemma 4 31B and Grok-3 are closely matched on the overall LLM Stats Score at 33.1 and 30.4.
In the 1 individual benchmarks reported for both models, Grok-3 wins 1; this is a narrower head-to-head signal than the composite indexes.
On price, Gemma 4 31B is roughly 39.3x cheaper per token on a blended 3:1 input/output basis, which adds up quickly at production volume.
Gemma 4 31B also accepts a larger context window (262,144 input tokens), making it the stronger choice for long documents and large codebases.
Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.
Choose Gemma 4 31B
- cost matters — it's about 39.3x cheaper per token
- you process long inputs — it offers a 262,144 token context window
- you want the most recent training data — it shipped Apr 2026
- you need open weights you can self-host or fine-tune
Choose Grok-3
- you value its reported benchmark strengths — it wins 1 of 1 exact shared results
At a glance
The differences that matter most.
Capability indexes
Additional strengths measured across groups of related public benchmarks
Individual benchmarks
12 reported for Gemma 4 31B · 5 for Grok-3
Gemma 4 31B outperforms in 0 benchmarks, while Grok-3 is better at 1 benchmark (GPQA).
Grok-3 significantly outperforms across most benchmarks.
Human preference
Blind head-to-head votes and playground preference scores
Pricing Analysis
Price comparison per million tokens
For input processing, Gemma 4 31B ($0.09/1M tokens) is 33.3x cheaper than Grok-3 ($3.00/1M tokens).
For output processing, Gemma 4 31B ($0.34/1M tokens) is 44.1x cheaper than Grok-3 ($15.00/1M tokens).
In conclusion, Grok-3 is more expensive than Gemma 4 31B.*
* Using a 3:1 ratio of input to output tokens
Context Window
Maximum input and output token capacity
Gemma 4 31B accepts 262,144 input tokens compared to Grok-3's 128,000 tokens. Gemma 4 31B can generate longer responses up to 262,144 tokens, while Grok-3 is limited to 8,000 tokens.
Input capabilities
Documented input modalities across available providers
Both Gemma 4 31B and Grok-3 support multimodal inputs.
They are both capable of processing various types of data, offering versatility in application.
Gemma 4 31B
Grok-3
License
Usage and distribution terms
Gemma 4 31B is licensed under Apache 2.0, while Grok-3 uses a proprietary license.
License differences may affect how you can use these models in commercial or open-source projects.
Apache 2.0
Open weights
Proprietary
Closed source
Release Timeline
When each model was launched
Gemma 4 31B was released on 2026-04-02, while Grok-3 was released on 2025-02-17.
Gemma 4 31B is 14 months newer than Grok-3.
Apr 2, 2026
5 months ago
1.1yr newerFeb 17, 2025
1.6 years ago
Knowledge Cutoff
When training data ends
Gemma 4 31B has a knowledge cutoff of 2025-01-01, while Grok-3 has a cutoff of 2024-11-17.
Gemma 4 31B has more recent training data (up to 2025-01-01), making it potentially better informed about events through that date compared to Grok-3 (2024-11-17).
Jan 2025
2 mo newerNov 2024
Provider Availability
Gemma 4 31B is available from DeepInfra, FriendliAI, Novita, Together. Grok-3 is available from xAI.
Gemma 4 31B
Grok-3
Outputs Comparison
Judge for yourself.
Run your own prompts against Gemma 4 31B and Grok-3 side-by-side, then vote on the output you prefer.
FAQ
Common questions about Gemma 4 31B vs Grok-3.