impulse

Gemma 4 26B A4B IT

Open weights

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind.

API id gemma-4-26b-a4b-itReleased 3 Apr 2026 (Apr 2026)
Input / 1M
$0.15
at Vertex AI
Output / 1M
$0.60
at Vertex AI
Context window
262K
tokens
Max output
262K
tokens per reply
Hosts
9
10 listings
Released
Apr 2026
3 Apr 2026

Pricing

List prices at Vertex AI, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.15$0.60
Cached input
90% off input
$0.015

What it costs a month

At Vertex AI list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$77.85
  • RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$61.20
  • Coding agent1,000 requests a day, 40K in / 2K out, cached$103
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$504
Use your own numbers

Where to run it

9 hosts serve Gemma 4 26B A4B IT. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
gemma-4-26b-a4b-it
$0.15 / $0.60
gemma-4-26b-a4b-it
-
DeepInfraInference
gemma-4-26B-A4B-it
$0.07 / $0.34
−48% vs headline
gemma-4-26b-a4b-it
$0.10 / $0.30
−43% vs headline
W&B InferenceInference
gemma-4-26B-A4B-it
$0.10 / $0.30
−43% vs headline
GMI CloudInference
gemma-4-26b-a4b-it
$0.13 / $0.40
−25% vs headline
Novita AIInference
gemma-4-26b-a4b-it
$0.13 / $0.40
−25% vs headline
AIHubMixGateway
gemma-4-26b-a4b-it
$0.14 / $0.40
−22% vs headline
gemma-4-26b-a4b-it
$0.25 / $0.50
+19% vs headline
Show 1 more listings: regions, snapshots and variants
Vertex AI
Alias
gemma-4-26b-a4b-it-maas
$0.15 / $0.60

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Open weights
License: Gemma License
Runs in about 61.9 GB of GPU memory
Runtimes: Ollama, llama.cpp, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

  • DeepInfra
    gemma-4-26B-A4B-it
    $0.07 / $0.34
    48% less
  • W&B Inference
    gemma-4-26B-A4B-it
    $0.10 / $0.30
    43% less
  • Cloudflare Workers AI
    gemma-4-26b-a4b-it
    $0.10 / $0.30
    43% less
  • GMI Cloud
    gemma-4-26b-a4b-it
    $0.13 / $0.40
    25% less
  • Novita AI
    gemma-4-26b-a4b-it
    $0.13 / $0.40
    25% less

Smaller Google models

No cheaper Google model keeps its key capabilities.

Other frontier labs

  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    24% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    24% less

Open-weight picks

  • Qwen3.8 27B
    Alibaba Qwen · on EmpirioLabs AI
    $0.17 / $0.50
    Has every key capability
    4% less
  • Ministral 14B 2512
    Mistral AI · on Mistral AI
    $0.20 / $0.20
    Has every key capability
    24% less
  • Nemotron 3.5 Super VL
    NVIDIA · on Kyma API
    $0.081 / $0.472
    Has every key capability
    32% less
  • Gemma 4 31B IT
    Google · on W&B Inference
    $0.10 / $0.34
    Has every key capability
    39% less

Call it

Using Vertex AI's API.

Python
from google import genai

client = genai.Client()  # reads GEMINI_API_KEY

response = client.models.generate_content(model="gemma-4-26b-a4b-it", contents="Hello")
print(response.text)

Quick answers

How much does Gemma 4 26B A4B IT cost?

On Vertex AI, Gemma 4 26B A4B IT costs $0.15 per million input tokens and $0.60 per million output tokens. Cached input costs $0.015 per million tokens. The cheapest host is DeepInfra, at $0.07 per million input tokens and $0.34 per million output tokens.

Which APIs offer Gemma 4 26B A4B IT?

Gemma 4 26B A4B IT runs on 9 APIs: Vertex AI, Google AI Studio, DeepInfra, Cloudflare Workers AI, W&B Inference, GMI Cloud, Novita AI, AIHubMix and Scaleway.

What is Gemma 4 26B A4B IT's context window?

262K tokens, with up to 262K tokens in one reply.

Is Gemma 4 26B A4B IT being retired?

No retirement date is scheduled for it on any API we track.

What are cheaper alternatives to Gemma 4 26B A4B IT?

Models that cost less and keep most of what it can do: Claude Haiku 5.5 on Anthropic (24% less), GPT-6 Luna on OpenAI (24% less) and Qwen3.8 27B on EmpirioLabs AI (4% less). Gemma 4 26B A4B IT itself costs 48% less on DeepInfra. Savings are on a mix of 3 input tokens to 1 output token.

When was Gemma 4 26B A4B IT released?

Google released it on 3 Apr 2026.

What is Gemma 4 26B A4B IT's API model id?

Use gemma-4-26b-a4b-it with Vertex AI. Other hosts use their own ids: gemma-4-26B-A4B-it on DeepInfra, gemma-4-26B-A4B-it on W&B Inference.

Sources

Raw catalog entry for gemma-4-26b-a4b-it on Vertex AI
{
  "cache_read_input_token_cost": 1.5e-8,
  "input_cost_per_token": 1.5e-7,
  "litellm_provider": "vertex_ai",
  "mode": "chat",
  "output_cost_per_token": 6e-7,
  "source": "https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing"
}