Textby Google

Gemma 4 26B A4B

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...

  • By price it is in the cheaper half of the 348 text models GenMagic runs (46th cheapest).
  • It has scores on 18 public benchmarks, listed below with their sources.
  • It was listed on Apr 3, 2026.

Data last changed Nov 10, 2026 · tracked since Oct 29, 2026 · Every model change

Try Gemma 4 26B A4B

Generate with Gemma 4 26B A4B right here. Verify your email to start with free credits.

<$0.01
Try

Specifications

Context window

262K tokens

Max output

236K tokens

Input price

$0.11 / 1M tokens

Output price

$0.38 / 1M tokens

Cached input

$0.064 / 1M tokens

Benchmarks

How Gemma 4 26B A4B scores on public benchmarks. Each card links to the benchmark explainer and to the original source.

2.9

source date 2026-11-11
Source: openrouter

39.3

source date 2026-11-11
Source: openrouter
AIMEMathematics

82.2%± 5.8

minimalsource date 2026-08-28
Source: epoch
Chess PuzzlesReasoning & Knowledge

6%± 2.4

minimalsource date 2026-08-06
Source: epoch
DTBenchReasoning & Knowledge

74.9%

Source: epoch
GPQA DiamondReasoning & Knowledge

70.2%± 4.7

source date 2026-10-31
Source: openrouter
LMArena (Coding)Instruction & Preference

1446 Elo

source date 2026-11-09
Source: lmarena
LMArena: Creative WritingInstruction & Preference

1402 Elo

source date 2026-11-09
Source: lmarena
LMArena: Hard PromptsInstruction & Preference

1439 Elo

source date 2026-11-09
Source: lmarena

1428 Elo

source date 2026-11-09
Source: lmarena
LMArena: MathMathematics

1476 Elo

source date 2026-11-09
Source: lmarena
LMArena: Multi-TurnInstruction & Preference

1442 Elo

source date 2026-11-09
Source: lmarena
LMArena (Overall)Instruction & Preference

1434 Elo

source date 2026-11-09
Source: lmarena
LMCAReasoning & Knowledge

29.7

Source: epoch
tau-bench (Airline)Agentic & Tool Use

68.7%± 4.5

source date 2026-10-31
Source: openrouter
WebDev ArenaAgentic & Tool Use

1361 Elo

Source: epoch
WeirdMLCoding

35.2%± 0.0

Source: epoch

Compare against every model in the full benchmark matrix.

Capabilities

Reasoning

Works through multi-step problems before answering, for harder tasks.

Vision input

Accepts images alongside text, so you can prompt with pictures and screenshots.

Tool use

Can call functions and tools, enabling agentic and structured workflows.

Structured output

Returns validated JSON that conforms to a schema you define.

Use Gemma 4 26B A4B through the API

The model id is google/gemma-4-26b-a4b-it. Requests bill your GenMagic balance at the price above, and each response reports what it cost.

curl
curl https://genmagic.co/api/v1/chat/completions \
  -H "Authorization: Bearer $GENMAGIC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemma-4-26b-a4b-it",
    "messages": [{ "role": "user", "content": "Explain vector databases in two sentences." }]
  }'
Python (openai SDK)
from openai import OpenAI

client = OpenAI(base_url="https://genmagic.co/api/v1", api_key="YOUR_GENMAGIC_API_KEY")

resp = client.chat.completions.create(
    model="google/gemma-4-26b-a4b-it",
    messages=[{"role": "user", "content": "Explain vector databases in two sentences."}],
)
print(resp.choices[0].message.content)

From an AI agent, call generate_text with "model": "google/gemma-4-26b-a4b-it" on the GenMagic MCP server. Keys, limits and every endpoint are in the API documentation.

Use Gemma 4 26B A4B in n8n

  1. In n8n, open the nodes panel, search for GenMagic and install it (a verified community node), then add it to your workflow.
  2. Set Resource to Text and Operation to Generate.
  3. In Model Name or ID, pick Google: Gemma 4 26B A4B , or enter google/gemma-4-26b-a4b-it as an expression.
  4. Connect your GenMagic API key (get one free). Each run is billed at the price above, and the node returns what it cost.

Ready-to-import workflows that already use the node are on GenMagic for n8n.

Similar-priced text models from other providers

Other text models GenMagic runs at a comparable price, on the same balance and API key.

ModelProviderPriceContext
Google: Gemma 4 26B A4B (this model)Google$0.11 in / $0.38 out per 1M tokens262K
Qwen: Qwen3.5-FlashQwen$0.098 in / $0.39 out per 1M tokens1M
Nex AGI: Nex-N2.5-ProNEX-AGI$0.11 in / $0.38 out per 1M tokens262K
DeepSeek: DeepSeek V4 Flash LatestDeepSeek$0.065 in / $0.45 out per 1M tokens1M
Meta: Llama Guard 4 12BMeta$0.27 in / $0.27 out per 1M tokens164K
Mistral: Ministral 3 8B 2512Mistral AI$0.23 in / $0.23 out per 1M tokens262K

More from Google

All 41 Google models

Price and listing history

Every change GenMagic has recorded for Gemma 4 26B A4B since it began tracking the model on Oct 29, 2026. Prices are GenMagic's, the same the studio and the API charge.

DateChangeBeforeAfter
Nov 10, 2026Price down 11%$0.14 in / $0.45 out per 1M tokens$0.12 in / $0.15 out per 1M tokens
Nov 10, 2026Updated: max output--
Nov 9, 2026Price up 33%$0.10 in / $0.34 out per 1M tokens$0.14 in / $0.45 out per 1M tokens
Nov 7, 2026Price down 25%$0.14 in / $0.45 out per 1M tokens$0.10 in / $0.34 out per 1M tokens
Nov 6, 2026Price up 260%$0.037 in / $0.11 out per 1M tokens$0.14 in / $0.45 out per 1M tokens
Nov 6, 2026Updated: max output--
Nov 5, 2026Price down 75%$0.15 in / $0.45 out per 1M tokens$0.037 in / $0.11 out per 1M tokens
Nov 4, 2026Price up 68%$0.088 in / $0.26 out per 1M tokens$0.15 in / $0.45 out per 1M tokens
Nov 3, 2026Price up 55%$0.057 in / $0.17 out per 1M tokens$0.088 in / $0.26 out per 1M tokens
Nov 2, 2026Price down 44%$0.10 in / $0.34 out per 1M tokens$0.057 in / $0.17 out per 1M tokens
Nov 2, 2026Updated: max output--
Nov 1, 2026Price up 36%$0.074 in / $0.25 out per 1M tokens$0.10 in / $0.34 out per 1M tokens
Oct 31, 2026Price down 25%$0.099 in / $0.33 out per 1M tokens$0.074 in / $0.25 out per 1M tokens

Where these figures come from

  • Price: GenMagic's own price, what the studio and the API charge, recomputed from the provider's price list every few minutes and recorded daily. How pricing works.
  • Specs and description: as Google publishes them (model card on Hugging Face), checked daily.
  • Benchmarks: 18 scores from Artificial Analysis and OpenRouter, Epoch AI, LMArena, re-read daily. Each score above links to its source.
  • Changes: the data on this page last changed on Nov 10, 2026, tracked since Oct 29, 2026. Every model's new listings, price changes and retirements are in the model changes log, and how every figure on GenMagic is sourced is on our sources page.

Generate with Gemma 4 26B A4B today

One GenMagic plan unlocks Gemma 4 26B A4B and every other model across text, image, audio, and video.