Gemma 4 26B A4B (free) API
Gemma 4 26B A4B is an open-weight multimodal Mixture-of-Experts model from Google DeepMind with 26B total and 4B active parameters, a 256K token context, and support for text, image, and video input.
Gemma 4 26B A4B is an open-weight multimodal model from Google DeepMind built on a Mixture-of-Experts architecture with 26B total parameters and 4B active parameters at inference. It accepts text, image, and video input and produces text output, with a 256K token context window. Images can have variable aspect ratios and resolutions, and video is processed up to 60 seconds at 1 frame per second. The model supports over 140 languages, native system prompts, native function calling, and a configurable thinking mode.
It is designed for coding assistance in IDEs and agentic workflows, code generation, debugging and refactoring, scientific and mathematical problem solving, research support, document parsing and OCR, image-based question answering, video understanding, and large-scale document processing and RAG systems. Its strengths include a balance of intelligence against parameter efficiency, near-frontier performance, and the ability to run on consumer GPUs and workstations. Reported throughput is about 91 tokens per second with a time to first token of about 0.92 seconds, and it is said to run within 2% of the 31B dense model on every benchmark tested.
Gemma 4 26B A4B is licensed under Apache 2.0 and available for commercial use. It can be accessed through Cloudflare, Google Cloud Vertex AI, NVIDIA NGC, and Azure ML, and API access is offered by 7 providers. It does not natively support audio; audio processing is available only in the smaller E2B, E4B, and 12B models. The model is positioned as the mid-range mainstay between the edge-oriented small models and the enterprise-oriented large model in the Gemma 4 family.
Updated:
MODEL ID
gemma-4-26b-a4b-it:free- Context
- 262.144K
- Max output
- 32.768K
- Recommended API
- messages
- Input
- image / text / video
- Output
- text
Highlights
- Open-weight multimodal Mixture-of-Experts model
- 26B total parameters, 4B active at inference
- 256K token context window
- Text, image, and video input
- Native function calling and thinking mode
- Apache 2.0 licensed for commercial use
Providers
Compare API prices across providers. Select a provider to view all charges and conditions.
| Uptime | |||
|---|---|---|---|
| $0 | $0 |
Historical observations are not a live health check. Missing data does not mean an outage.
Benchmarks
- 82.6%
MMLU Pro
Score on the MMLU Pro benchmark.
- 88.3%
AIME 2026 (no tools)
Score on AIME 2026 without tools.
- 77.1%
LiveCodeBench v6
Score on LiveCodeBench v6.
- 82.3%
GPQA Diamond
Score on GPQA Diamond.
- 1718
Codeforces ELO
Elo rating on Codeforces.
- 73.8%
MMMU Pro
Score on MMMU Pro.
- 82.4%
MATH-Vision
Score on MATH-Vision.
- 44.1%
MRCR v2 8 needle 128k
Result on the long-context MRCR v2 8-needle 128k evaluation.
Scores as published by the model provider.
Reasoning
- Enabled by default
- No
- Always enabled
- No
Quick Start
Drop-in code to call this model with Felo API Platform compatible APIs. Start in the Playground when you want to test prompts before writing code.
Get your API key
Create a key from your Felo API Platform account and set it as an environment variable.
export FELO_API_KEY="YOUR_API_KEY"Make your first request
Use this model ID with the Felo API Platform endpoint. Optional request headers such as X-Request-Id can help troubleshooting.
gemma-4-26b-a4b-it:freeThis model is reasoning-capable. If your selected protocol and model route expose reasoning controls, preserve the returned reasoning metadata when continuing a conversation.
const response = await fetch("https://openapi.felo.ai/api/v1/messages", {
method: 'POST',
headers: {
Authorization: `Bearer ${process.env.FELO_API_KEY}`,
'Content-Type': 'application/json',
},
body: JSON.stringify({
"model": "gemma-4-26b-a4b-it:free",
"max_tokens": 1024,
"messages": [
{
"role": "user",
"content": "Hello"
}
]
}),
});
if (!response.ok) throw new Error(await response.text());
console.log(await response.json());Using third-party SDKs
Use compatible SDKs by setting baseURL to https://openapi.felo.ai/api/v1. Use the Anthropic-compatible base URL for Claude-style clients. https://openapi.felo.ai/api.
Enable streaming
Add stream: true to receive responses as server-sent events.
curl -N https://openapi.felo.ai/api/v1/messages \
-H "Authorization: Bearer $FELO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gemma-4-26b-a4b-it:free",
"max_tokens": 1024,
"messages": [
{
"role": "user",
"content": "Hello"
}
],
"stream": true
}'Endpoint
Felo API Platform exposes compatible model-calling surfaces. Pick the surface your app or agent client already supports.
Messages
Anthropic-compatible messages for Claude-style clients.
POST
https://openapi.felo.ai/api/v1/messagesParameters
modelstringgemma-4-26b-a4b-it:freePublic model ID to call.max_tokensinteger—Maximum generated tokens.messagesarray—Conversation messages for Chat Completions.seed———toolsarray—Tool calling definitions following compatible schema.top_p———reasoning———temperaturenumber—Controls response variety when supported.tool_choicestring or object—Controls whether a tool should be called.response_formatobject—Request structured output when supported.include_reasoning———Frequently asked questions
What is Gemma 4 26B A4B (free)?
Gemma 4 26B A4B is an open-weight multimodal model from Google DeepMind built on a Mixture-of-Experts architecture with 26B total parameters and 4B active parameters at inference. It accepts text, image, and video input and produces text output, with a 256K token context window. Images can have variable aspect ratios and resolutions, and video is processed up to 60 seconds at 1 frame per second. The model supports over 140 languages, native system prompts, native function calling, and a configurable thinking mode. It is designed for coding assistance in IDEs and agentic workflows, code generation, debugging and refactoring, scientific and mathematical problem solving, research support, document parsing and OCR, image-based question answering, video understanding, and large-scale document processing and RAG systems. Its strengths include a balance of intelligence against parameter efficiency, near-frontier performance, and the ability to run on consumer GPUs and workstations. Reported throughput is about 91 tokens per second with a time to first token of about 0.92 seconds, and it is said to run within 2% of the 31B dense model on every benchmark tested. Gemma 4 26B A4B is licensed under Apache 2.0 and available for commercial use. It can be accessed through Cloudflare, Google Cloud Vertex AI, NVIDIA NGC, and Azure ML, and API access is offered by 7 providers. It does not natively support audio; audio processing is available only in the smaller E2B, E4B, and 12B models. The model is positioned as the mid-range mainstay between the edge-oriented small models and the enterprise-oriented large model in the Gemma 4 family.
How is usage billed?
Your bill depends on the provider you use, its prices and your usage. Longer inputs or other conditions may have different prices.
What input modalities does Gemma 4 26B A4B support?
It accepts text, image, and video input and produces text output. Images can have variable aspect ratios and resolutions, and video is processed up to 60 seconds at 1 frame per second.
How large is the context window?
The model has a 256K token context window. The provided materials do not specify a separate output token limit.
What tasks is it best suited for?
It is suited for coding assistance and agentic workflows, code generation, debugging and refactoring, scientific and mathematical problem solving, research support, document parsing and OCR, image-based question answering, video understanding, and large-scale document processing and RAG.
Does it support audio input?
No. It does not natively support audio; audio processing is available only in the smaller E2B, E4B, and 12B models.
How does it compare with the other Gemma 4 models?
It sits between the edge-oriented small models (E2B, E4B) and the enterprise-oriented 31B model. It is reported to run within 2% of the 31B model on every benchmark tested, while scoring slightly below it.
Can it be used commercially?
Yes. It is licensed under Apache 2.0 and available for commercial use.
Related models
Gemini 4 Carbon
Googlegemini-4-carbonGemini 4 Carbon is Google's strongest frontier model, built on Gemini 4 Argon for long-horizon agentic workflows, cybersecurity, scientific reasoning, and complex knowledge work.
Context
1.049M
Max output
943.718K
Input tokens
$4 / 1M tokens
Output tokens
$20 / 1M tokens
Gemini 4 Carbon (batch)
Googlegemini-4-carbon:batchGemini 4 Carbon is Google's multimodal frontier model for long-horizon agentic work, with 1 million token context and adjustable reasoning effort.
Context
1.049M
Max output
943.718K
Input tokens
$2 / 1M tokens
Output tokens
$10 / 1M tokens
Gemini 4 Argon
Googlegemini-4-argonGoogle's Gemini 4 Argon is a frontier model with a 1,048,576-token context window, multimodal input support, and top scores on 14 of 18 internal benchmarks.
Context
1.049M
Max output
943.718K
Input tokens
$4 / 1M tokens
Output tokens
$20 / 1M tokens
Gemini 3.8 Flash
Googlegemini-3.8-flashGemini 3.8 Flash is Google's most intelligent Flash model, a reasoning model with 1,000,000-token context and multimodal input for long-horizon agents and enterprise workflows.
Context
1.049M
Max output
65.536K
Input tokens
$1.5 / 1M tokens
Output tokens
$7.5 / 1M tokens
Gemini 2.5 Pro
Googlegemini-2.5-proGoogle DeepMind's Gemini 2.5 Pro is a hybrid-reasoning, multimodal model with a 1 million token context window, built for advanced coding, agentic applications, and video understanding.
Context
1.049M
Max output
65.536K
Input tokens · Base rate
$1.25 / 1M tokens
Output tokens · Base rate
$10 / 1M tokens