vs
0:7
overall capability wins (2 ties across 9 scored). Gemma 4 31B leads overall.
- Reasoning0:3
- Coding0:2
- Arena0:1
- Pricing0:3
- Speed2:0
- Specs0:1
Verdict
Gemma 4 31B leads coding (widest gap: +2.9 on LiveCode); Gemma 4 31B ranks higher on LMArena (+13 Elo); Gemma 4 31B is ~1.7x cheaper on a blended token basis than Gemma 4 26B.
- Gemma 4 31B leads coding (widest gap: +2.9 on LiveCode)
- Gemma 4 31B ranks higher on LMArena (+13 Elo)
- Gemma 4 31B is ~1.7x cheaper on a blended token basis than Gemma 4 26B
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Gemma 4 26B
- $0.30
- Gemma 4 31B
- $0.17
Gemma 4 31B is about 1.7x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Family
Gemma
Gemma
License
Apache 2.0
Apache 2.0
Open weights
Yes
Yes
Release
Apr 8, 2026
Apr 8, 2026
Knowledge cutoff
-
-
API / provider
Google AI / self-host
Google AI / self-host
Modalities
text, image → text
text, image → text
Specs
Context window
256K
256K
Max output
8K
8K
Parameters
26B (3.8B act.)
31B
Pricing
Input $/1M
$0.15
$0.10
Output $/1M
$0.60
$0.30
Blended $/1M (3∶1)
$0.26
$0.15
Speed
tok/s
95
85
TTFT (s)
0.28
0.3
Reasoning
MMLU-Pro
82.6
85.2
GPQA Diamond
82.3
85.7
Humanity's Last Exam
8.7
23.6
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
20.0
AA-LCR v1.1
—
69.7
CritPt
—
1.4
MMMU-Pro
—
73.4
IFBench
—
75.6
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
77.1
80.0
Terminal-Bench 2.1
—
43.4
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
45.5
CursorBench
—
—
SWE-Rebench
—
25.1
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,362
1,364
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,438
1,451
| Metric | Gemma 4 26B | Gemma 4 31B | Delta |
|---|---|---|---|
| Identity | |||
| Organization | — | ||
| Family | Gemma | Gemma | — |
| License | Apache 2.0 | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Apr 8, 2026 | Apr 8, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Google AI / self-host | Google AI / self-host | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 256K | 256K | tie |
| Max output | 8K | 8K | tie |
| Parameters | 26B (3.8B act.) | 31B | B +5.00 |
| Pricing | |||
| Input $/1M | $0.15 | $0.10 | B +$0.05 |
| Output $/1M | $0.60 | $0.30 | B +$0.30 |
| Blended $/1M (3∶1) | $0.26 | $0.15 | B +$0.11 |
| Speed | |||
| tok/s | 95 | 85 | A +10 |
| TTFT (s) | 0.28 | 0.3 | A +0.02 |
| Reasoning | |||
| MMLU-Pro | 82.6 | 85.2 | B +2.6 pts |
| GPQA Diamond | 82.3 | 85.7 | B +3.4 pts |
| Humanity's Last Exam | 8.7 | 23.6 | B +14.9 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 20.0 | — |
| AA-LCR v1.1 | — | 69.7 | — |
| CritPt | — | 1.4 | — |
| MMMU-Pro | — | 73.4 | — |
| IFBench | — | 75.6 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 77.1 | 80.0 | B +2.9 pts |
| Terminal-Bench 2.1 | — | 43.4 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 45.5 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | 25.1 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,362 | 1,364 | B +2 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,438 | 1,451 | B +13 Elo |
Gemma 4 26B
- Context
- 256K / 8K out
- Parameters
- 26B (3.8B act.)
- Price
- $0.15 / $0.60
- Speed
- 95 tok/s · 0.28s TTFT
- Modalities
- text, image → text
- License
- Apache 2.0
Open Gemma 4 mid-size instruct model available via Google Cloud Agent Platform and self-host.
Gemma 4 31B
- Context
- 256K / 8K out
- Parameters
- 31B
- Price
- $0.10 / $0.30
- Speed
- 85 tok/s · 0.3s TTFT
- Modalities
- text, image → text
- License
- Apache 2.0
Open-weight Gemma 4 instruction model - strong mid-size open alternative under Apache 2.0.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.