Gemma 4 Benchmarks
with source transparency
This page separates release-day ranking claims from the latest available leaderboard snapshot so ranking statements stay date-scoped and auditable.
Core benchmark table
Primary rows are sourced from official model documentation where available.
| Benchmark | Description | Gemma 4 31B | 26B A4B | E4B | E2B | Gemma 3 27B |
|---|---|---|---|---|---|---|
| AIME 2026 | Competition math reasoning | 88.3% Official · confidence high Published by Hugging Face with Google-provided benchmark table. | Pending official publication | Pending official publication | ||
| tau2-bench | Agentic tool-use accuracy | 68.2% Official · confidence high Value from HF benchmark table (Tau2 average over 3). | Pending official publication | Pending official publication | ||
| Arena AI ELO | Text arena conversation quality | 1452 (release claim) Official · confidence high Google launch post (Apr 2, 2026) states 31B as #3 open model. | 1441 (release claim) Official · confidence high HF launch blog states 26B A4B estimated arena score 1441. | Pending official publication | Pending official publication | Pending official publication |
| OmniDocBench 1.5 | Document OCR/edit distance (lower is better) | Pending official publication | Pending official publication |
Cross-generation uplift
AIME 2026
+328.8% uplift
tau2-bench
+1209.1% uplift
Arena ranking context
We keep two timelines visible: launch-day claims (Apr 2, 2026 post, data as of Apr 1) and a later leaderboard snapshot to avoid stale absolute ranking claims.
Gemma 4 31B Dense
ELO 1452
#3 open model
Snapshot: 2026-04-01 · checked 2026-04-11
Google Keyword - Gemma 4: Byte for byte, the most capable open modelsGemma 4 26B MoE
ELO 1441
#6 open model
Snapshot: 2026-04-01 · checked 2026-04-11
Google Keyword - Gemma 4: Byte for byte, the most capable open models| Model | Arena AI ELO | Global Rank | Open Rank | License | Context |
|---|---|---|---|---|---|
GLM-5.1 n/a | 1458.51 | #26 overall | #1 open | Open (commercial limits vary by provider) | ~203K |
GLM-5 n/a | 1457.14 | #27 overall | #2 open | Open (commercial limits vary by provider) | ~200K |
Kimi 2.5 n/a | 1452.60 | #28 overall | #3 open | Open (commercial limits vary by provider) | ~128K |
Gemma 4 31B Dense 31B dense | 1451.16 | #29 overall | #4 open | Apache 2.0 | 256K |
Qwen 3.5 32B 32B | 1445.18 | #34 overall | #5 open | Apache 2.0 | ~128K |
GLM-4.7 n/a | 1442.65 | #37 overall | #6 open | Open (commercial limits vary by provider) | n/a |
Gemma 4 26B MoE 26B MoE (4B active) | 1437.89 | #43 overall | #7 open | Apache 2.0 | 256K |
Latest snapshot date in this table: 2026-04-07 · checked 2026-04-11
Source: Hugging Face - lmarena-ai/leaderboard-dataset (text_style_control/latest)
Ranking can drift between snapshots. Keep date qualifiers in any homepage, model, or blog claim.
Competitive watch: gemma4.wiki has started publishing benchmark-oriented pages. We are tracking this as a content velocity signal.
Sources used on this page
Google Keyword - Gemma 4: Byte for byte, the most capable open models
official · checked 2026-04-11
Google AI for Developers - Gemma 4 Model Card
official · checked 2026-04-11
Google DeepMind - Gemma 4
official · checked 2026-04-11
Arena AI - Text Arena Leaderboard
official · checked 2026-04-11
Hugging Face - lmarena-ai/leaderboard-dataset (text_style_control/latest)
official · checked 2026-04-11
Google AI for Developers - Prompt Formatting
official · checked 2026-04-08
Google AI for Developers - Function Calling
official · checked 2026-04-08
vLLM Docs - Gemma 4
official · checked 2026-04-08
Unsloth Docs - Gemma 4 Fine-tuning
official · checked 2026-04-08
Hugging Face Blog - Gemma 4
media · checked 2026-04-11
Hugging Face - onnx-community/gemma-4-E2B-it-ONNX
official · checked 2026-04-11
gemma4.wiki - Gemma 4 Models
community · checked 2026-04-11
VentureBeat - Gemma 4 coverage
media · checked 2026-04-08
DEV Community - Gemma 4 deployment guide
community · checked 2026-04-08
Spheron Network - Gemma 4 deployment
community · checked 2026-04-08
Hacker News - Gemma 4 hardware reports
community · checked 2026-04-08
AMD Day-0 support for Gemma 4
official · checked 2026-04-08