Home/Models
Models/Google

Gemma 4 26B A4B IT

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 a fraction of the compute cost. Supports multimodal input including text, images, and video (up to 60s at 1fps). Features a 256K token context window, native function calling, configurable thinking/reasoning mode, and structured output support. Released under Apache 2.0.

google/gemma-4-26b-a4b-it
Organization
Google
Family
gemma
Providers
28
Context
262,144
Output limit
32,768
Knowledge
—
Release
2026-04-02
Updated
—
Weights
Open
Input
text, image/png, image/jpeg, image/webp, image/heic, image/heif
Output
text
Capabilities
—

Providers

Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.

43 offers · USD per million tokens
ProviderModel IDContextOutputInput / output · 1MReasoningToolsStructuredDetails
AI/ML API
google/gemma-4-26b-a4b-it
262,144128,000— / ————
View
AKI.IO
gemma4-26b
256,00032,768— / ————
View
Amazon Bedrock
google.gemma-4-26b-a4b
262,14432,768— / ————
View
Charm Hyper
gemma-4-26b-a4b-it
256,00025,600— / ————
View
Cloudflare Workers AI
@cf/google/gemma-4-26b-a4b-it
256,00016,384— / ————
View
CoreWeave
google/gemma-4-26B-A4B-it
262,144262,144— / ————
View
Cortecs
gemma-4-26b-a4b-it
262,00081,920— / ————
View
Deep Infra
google/gemma-4-26B-A4B-it
262,14432,768— / ————
View
DevPass (LLM Gateway)
gemma-4-26b-a4b-it
262,14432,768— / ————
View
EmpirioLabs AI
gemma-4-26b-a4b
262,14432,768— / ————
View
evroc
google/gemma-4-26B-A4B-it
262,14432,768— / ————
View
GreenPT
gemma4
262,14432,768— / ————
View
Hugging Face
google/gemma-4-26B-A4B-it
262,14432,768— / ————
View
Kilo Gateway
google/gemma-4-26b-a4b-it
262,144235,929— / ————
View
LLM Gateway
novita/gemma-4-26b-a4b-it
262,14432,768— / ————
View
LLM Gateway
deepinfra/gemma-4-26b-a4b-it
262,14432,768— / ————
View
Merge Gateway
google/gemma-4-26b-a4b-it
262,14465,536— / ————
View
NaN
gemma4
262,14432,768— / ————
View
NanoGPT
google/gemma-4-26b-a4b-it
262,144131,072— / ————
View
NanoGPT
google/gemma-4-26b-a4b-it:thinking
262,144131,072— / ————
View
Novita AI
google/gemma-4-26b-a4b-it
262,144131,072— / ————
View
Cloudflarevia OpenRouter
google/gemma-4-26b-a4b-it
unknown		256,000230,400$0.10 / $0.30YesYesNo
View
CoreWeavevia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,144235,929$0.10 / $0.30YesNoYes
View
Darkbloomvia OpenRouter
google/gemma-4-26b-a4b-it
unknown		131,07232,768$0.042 / $0.22YesYesYes
View
DeepInfravia OpenRouter
google/gemma-4-26b-a4b-it
fp8		262,14416,384$0.07 / $0.34YesYesYes
View
DekaLLMvia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,144235,929$0.06 / $0.33YesYesNo
View
Googlevia OpenRouter
google/gemma-4-26b-a4b-it
unknown		262,144235,929$0.15 / $0.60YesYesYes
View
Google AI Studiovia OpenRouter
google/gemma-4-26b-a4b-it:free
unknown		262,14432,768$0.00 / $0.00YesYesNo
View
Io Netvia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,142235,927$0.0483 / $0.21YesYesNo
View
Makoravia OpenRouter
google/gemma-4-26b-a4b-it
unknown		256,000128,000$0.08 / $0.32YesYesNo
View
NextBitvia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,144235,929$0.0675 / $0.225YesYesYes
View
Novitavia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,144131,072$0.13 / $0.40YesYesNo
View
Parasailvia OpenRouter
google/gemma-4-26b-a4b-it
bf16		262,144235,929$0.13 / $0.40YesNoYes
View
SiliconFlowvia OpenRouter
google/gemma-4-26b-a4b-it
fp8		262,144235,929$0.14 / $0.40YesNoYes
View
Venicevia OpenRouter
google/gemma-4-26b-a4b-it
bf16		256,0008,192$0.13 / $0.40YesYesYes
View
OrcaRouter
google/gemma-4-26b-a4b-it
262,14432,768— / ————
View
Requesty
gemma-4-26b-a4b-it
262,14432,768— / ————
View
Scaleway
gemma-4-26b-a4b-it
256,00016,384— / ————
View
SiliconFlow
google/gemma-4-26B-A4B-it
262,144262,144— / ————
View
Tempr Gateway
google/gemma-4-26b-a4b-it
262,14432,768— / ————
View
Venice AI
google-gemma-4-26b-a4b-it
256,0008,192— / ————
View
Vercel AI Gateway
google/gemma-4-26b-a4b-it
262,144131,072— / ————
View
Vertex
google/gemma-4-26b-a4b-it-maas
262,144128,000— / ————
View

Prices and limits apply to each offer. A dash means the value is not provided. OpenRouter offers show the hosting provider and routing channel separately.

Pricing

Compare token rates and cache charges across available routes and hosting configurations.

USD per million tokens
Input from$0.00Per 1M input tokens
Output from$0.00Per 1M output tokens
Available offers43Across 28 providers

Lowest listed input and output prices across the offers above; they may belong to different providers. These are base rates, before conditional pricing, discounts or additional fees.

Endpoint price comparisonLatest observed snapshot · 10 lowest-priced configurations
InputOutput
Google AI Studiogoogle-ai-studio
$0.00$0.00
Darkbloomdarkbloom
$0.042$0.22
Io Netio-net/bf16
$0.0483$0.21
DekaLLMdekallm/bf16
$0.06$0.33
NextBitnextbit/bf16
$0.0675$0.225
DeepInfradeepinfra/fp8
$0.07$0.34
Makoramakora
$0.08$0.32
Cloudflarecloudflare
$0.10$0.30
CoreWeavecoreweave/bf16
$0.10$0.30
Novitanovita/bf16
$0.13$0.40
google/gemma-4-26b-a4b-itObserved
Input
$0.0675
Output
$0.225
Cache read
$0.0375
google/gemma-4-26b-a4b-it:freeObserved
Input
$0.00
Output
$0.00
Hosting prices through OpenRouter All endpoints and cache rates
EndpointInput / 1MOutput / 1MCache read / 1MCache write / 1M
Cloudflarecloudflare$0.10$0.30$0.05—
CoreWeavecoreweave/bf16$0.10$0.30$0.05—
Darkbloomdarkbloom$0.042$0.22$0.021—
DeepInfradeepinfra/fp8$0.07$0.34——
DekaLLMdekallm/bf16$0.06$0.33$0.04—
Googlegoogle-vertex/global$0.15$0.60——
Io Netio-net/bf16$0.0483$0.21$0.0203—
Makoramakora$0.08$0.32$0.032—
NextBitnextbit/bf16$0.0675$0.225$0.0375—
Novitanovita/bf16$0.13$0.40——
Parasailparasail/bf16$0.13$0.40$0.05—
SiliconFlowsiliconflow/fp8$0.14$0.40$0.05—
Venicevenice/bf16$0.13$0.40$0.05—
Google AI Studiogoogle-ai-studio$0.00$0.00——

Each row is a hosting configuration. Open an offer in Providers for its full pricing conditions and additional fees. Endpoint observations may differ from the routing catalog's base price.

Performance

Review reported latency and output-throughput percentiles from the latest observed rolling window.

Latest 30-minute observation

Time to first token · milliseconds

Latency distributionRolling 30-minute snapshot · 6 endpoints
P50P75P90P99
Cloudflarecloudflare
539–2,487.45 ms
CoreWeavecoreweave/bf16
261–1,655.33 ms
Darkbloomdarkbloom
899–5,781.95 ms
DeepInfradeepinfra/fp8
514–2,269.45 ms
DekaLLMdekallm/bf16
500–3,982.35 ms
Googlegoogle-vertex/global
1,569.5–9,059.45 ms
All latency percentiles 14 endpoints
EndpointP50P75P90P99
Cloudflarecloudflare539 ms810 ms1,167.9 ms2,487.45 ms
CoreWeavecoreweave/bf16261 ms394 ms560 ms1,655.33 ms
Darkbloomdarkbloom899 ms1,521.25 ms2,404.8 ms5,781.95 ms
DeepInfradeepinfra/fp8514 ms754 ms1,039.9 ms2,269.45 ms
DekaLLMdekallm/bf16500 ms895 ms1,450.9 ms3,982.35 ms
Googlegoogle-vertex/global1,569.5 ms2,592 ms3,936 ms9,059.45 ms
Io Netio-net/bf16535 ms782 ms1,415 ms6,286.34 ms
Makoramakora263 ms407 ms751.7 ms2,051.54 ms
NextBitnextbit/bf16247 ms474.25 ms871 ms1,566.18 ms
Novitanovita/bf16757 ms1,065 ms1,459 ms2,862.26 ms
Parasailparasail/bf16689.5 ms995 ms1,405.8 ms4,765.07 ms
SiliconFlowsiliconflow/fp81,356.5 ms1,839 ms2,424.3 ms4,638.92 ms
Venicevenice/bf16899 ms1,459.5 ms2,199.9 ms6,094.37 ms
Google AI Studiogoogle-ai-studio1,041 ms1,338 ms2,449.8 ms10,271.88 ms

Output throughput · tokens per second

Throughput distributionRolling 30-minute snapshot · 6 endpoints
P50P75P90P99
Cloudflarecloudflare
48–170 t/s
CoreWeavecoreweave/bf16
81–149.27 t/s
Darkbloomdarkbloom
50–109 t/s
DeepInfradeepinfra/fp8
78–124 t/s
DekaLLMdekallm/bf16
74–187.26 t/s
Googlegoogle-vertex/global
31–82.58 t/s
All throughput percentiles 14 endpoints
EndpointP50P75P90P99
Cloudflarecloudflare48 t/s58 t/s104 t/s170 t/s
CoreWeavecoreweave/bf1681 t/s103 t/s120 t/s149.27 t/s
Darkbloomdarkbloom50 t/s68 t/s84 t/s109 t/s
DeepInfradeepinfra/fp878 t/s92 t/s104 t/s124 t/s
DekaLLMdekallm/bf1674 t/s97 t/s122 t/s187.26 t/s
Googlegoogle-vertex/global31 t/s40 t/s52 t/s82.58 t/s
Io Netio-net/bf1659 t/s74 t/s89 t/s120 t/s
Makoramakora212 t/s275 t/s342 t/s506.09 t/s
NextBitnextbit/bf1680 t/s92 t/s105 t/s125 t/s
Novitanovita/bf16106 t/s164 t/s196.3 t/s227 t/s
Parasailparasail/bf1624 t/s42 t/s71 t/s147.43 t/s
SiliconFlowsiliconflow/fp843 t/s58 t/s69 t/s89.81 t/s
Venicevenice/bf1655 t/s80 t/s93 t/s113.52 t/s
Google AI Studiogoogle-ai-studio40 t/s44 t/s45 t/s47 t/s

Percentiles describe the observed request distribution. These are snapshots of a rolling window, not a historical time series.

Uptime

Compare successful request rates across the latest 5-minute, 30-minute, and 24-hour observations.

Successful requests by window
Availability by endpointCurrent overlapping windows · 14 endpoints
Endpoint5 minutes30 minutes24 hours
Cloudflarecloudflare
CoreWeavecoreweave/bf16
Darkbloomdarkbloom
DeepInfradeepinfra/fp8
DekaLLMdekallm/bf16
Googlegoogle-vertex/global
Io Netio-net/bf16
Makoramakora
NextBitnextbit/bf16
Novitanovita/bf16
Parasailparasail/bf16
SiliconFlowsiliconflow/fp8
Venicevenice/bf16
Google AI Studiogoogle-ai-studio

Reported by OpenRouter at the time of collection; rate-limited requests are excluded. A dash means no measurement was reported. This is not a live availability check.

Benchmarks

Review attributed evaluation results on each source's original scale.

External evaluations
google/gemma-4-26b-a4b-itObserved
Benchmark profileArtificial Analysis · shared source scale from 0 to 40
Coding39.3
Agentic2.9
Intelligence16.7
google/gemma-4-26b-a4b-it:freeObserved
Benchmark profileArtificial Analysis · shared source scale from 0 to 40
Coding39.3
Agentic2.9
Intelligence16.7

Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.

Detailed evaluations

Evaluation scoresAccuracy or item F1 · higher is better
tau bench verified airlineGoogle: Gemma 4 26B A4B
69.5%
gpqa diamondGoogle: Gemma 4 26B A4B
70.647%
EvaluationConfigurationSourceResultsDetails
Composite indicesGemma 4 26B A4B (Reasoning)google/gemma-4-26b-a4b-it-20260403Artificial AnalysisIntelligence 16.7 · Coding 39.3 · Agentic 2.9
Details
tau bench verified airlineGoogle: Gemma 4 26B A4Bgoogle/gemma-4-26b-a4b-it-20260403OpenRouter69.5%
Details
gpqa diamondGoogle: Gemma 4 26B A4Bgoogle/gemma-4-26b-a4b-it-20260403OpenRouter70.647%
Details

Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 24, 2026.

Apps & session costs

Compare observed median session costs for applications using this model, grouped by conversation length.

30-day sample · through Oct 18, 2026
ApplicationModel routeTurnsMedian cost / sessionDetails
Hermes Agentgoogle/gemma-4-26b-a4b-it-2026040310–49 turns
$0.027469
Details
Hermes Agentgoogle/gemma-4-26b-a4b-it-202604031 turn
$0.000522
Details
Hermes Agentgoogle/gemma-4-26b-a4b-it-202604032–9 turns
$0.002173
Details

Source: OpenRouter, as of Oct 20, 2026. Published under CC BY 4.0. These are observed costs per session, not token prices or a forecast for your workload. Apps are compared separately.

Activity

Daily prompt and completion tokens reported for this model's routes in OpenRouter's top 50. Missing days have no published value.

Last 90 complete UTC days
Daily token usage3,640,318,008,861 reported tokens · 73 days with observations
Jul 26, 2026Oct 23, 2026
Daily observations · 73 days
Date (UTC)Reported tokensReported routes
Oct 22, 202656,180,203,8681
Oct 21, 202655,394,418,5861
Oct 20, 202659,717,456,9111
Oct 19, 202655,305,536,2721
Oct 18, 202645,636,724,5381
Oct 17, 202646,366,201,9961
Oct 15, 202656,095,684,4551
Oct 13, 202660,518,751,8351
Oct 12, 202661,067,289,3021
Oct 11, 202646,757,313,9091
Oct 10, 202644,891,993,5091
Oct 8, 202662,776,066,8621
Oct 7, 202660,090,604,9431
Oct 6, 202651,680,183,7141
Oct 5, 202652,959,118,1031
Sep 28, 202650,403,161,2531
Sep 27, 202642,454,663,3221
Sep 25, 202650,357,319,0741
Sep 24, 202651,962,449,2381
Sep 22, 202656,336,004,8421
Sep 21, 202649,772,338,9781
Sep 20, 202638,328,806,5421
Sep 19, 202636,742,569,1981
Sep 17, 202646,047,919,5661
Sep 15, 202642,888,051,7131
Sep 13, 202635,902,165,9311
Sep 11, 202644,827,162,5471
Sep 10, 202655,244,402,1341
Sep 9, 202645,212,397,3581
Sep 8, 202644,620,791,7821
Sep 7, 202649,154,356,7261
Sep 6, 202650,951,812,1081
Sep 5, 202648,050,107,0901
Sep 4, 202652,385,661,5551
Sep 3, 202657,475,159,9441
Sep 2, 202653,587,770,2001
Sep 1, 202648,026,919,0341
Aug 31, 202643,651,687,0711
Aug 30, 202634,210,091,0251
Aug 29, 202635,637,610,9491
Aug 28, 202639,779,623,9901
Aug 27, 202644,704,075,0601
Aug 26, 202653,331,209,5031
Aug 25, 202648,527,627,6561
Aug 24, 202647,361,988,4211
Aug 23, 202648,022,912,8481
Aug 22, 202645,469,111,2801
Aug 21, 202647,153,821,9171
Aug 20, 202649,208,552,2641
Aug 19, 202652,441,138,5131
Aug 18, 202658,684,560,2841
Aug 17, 202651,353,420,8991
Aug 16, 202650,853,260,5831
Aug 15, 202647,489,779,3031
Aug 14, 202649,563,625,3181
Aug 13, 202652,357,034,7131
Aug 12, 202649,905,803,3501
Aug 11, 202661,384,692,0781
Aug 10, 202660,430,213,0761
Aug 9, 202657,228,554,4961
Aug 8, 202659,781,700,6301
Aug 7, 202650,092,326,6881
Aug 6, 202646,048,533,7231
Aug 5, 202645,340,977,6871
Aug 4, 202642,479,379,1551
Aug 3, 202644,674,393,6851
Aug 2, 202646,116,286,6281
Aug 1, 202642,866,305,1111
Jul 31, 202653,614,101,2561
Jul 30, 202655,892,929,8891
Jul 29, 202656,677,753,4971
Jul 28, 202651,368,108,2551
Jul 27, 202650,445,279,1221

Source: OpenRouter (openrouter.ai/rankings), as of Oct 24, 2026. CC BY 4.0. Totals include only individually published routes. Tokenizers differ by provider; the aggregated “other” category is never assigned to a model.

Technical details & API

Explore specifications, supported parameters, reasoning controls, and API identifiers for each OpenRouter route.

Current route specifications
google/gemma-4-26b-a4b-itObserved
Context
262,144
Maximum output
235,929
Tokenizer
Gemma
Input
image, text, video
Output
text
Moderation
No
Added to OpenRouter
2026-04-03
Base URLhttps://openrouter.ai/api/v1
Model IDgoogle/gemma-4-26b-a4b-it
Version IDgoogle/gemma-4-26b-a4b-it-20260403
Weights IDgoogle/gemma-4-26B-A4B-it

Reasoning

Required
No
Enabled by default
No
Default effort
—
Supported efforts
—

Supported parameters

  • frequency_penalty
  • include_reasoning
  • logit_bias
  • logprobs
  • max_tokens
  • min_p
  • presence_penalty
  • reasoning
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_k
  • top_logprobs
  • top_p

Default parameters

top k
64
top p
0.95
temperature
1
google/gemma-4-26b-a4b-it:freeObserved
Context
262,144
Maximum output
32,768
Tokenizer
Gemma
Input
image, text, video
Output
text
Moderation
No
Added to OpenRouter
2026-04-03
Base URLhttps://openrouter.ai/api/v1
Model IDgoogle/gemma-4-26b-a4b-it:free
Version IDgoogle/gemma-4-26b-a4b-it-20260403
Weights IDgoogle/gemma-4-26B-A4B-it

Reasoning

Required
No
Enabled by default
No
Default effort
—
Supported efforts
—

Supported parameters

  • include_reasoning
  • max_tokens
  • reasoning
  • response_format
  • seed
  • temperature
  • tool_choice
  • tools
  • top_p

Default parameters

top k
64
top p
0.95
temperature
1

Frequently asked questions

Answers based on the model metadata and serving offers currently in the catalog.

How much does Gemma 4 26B A4B IT cost?

Listed rates start at $0.00 per million input tokens and $0.00 per million output tokens. Prices vary by provider and configuration; see Pricing for details.

Which providers offer Gemma 4 26B A4B IT?

There are 43 cataloged offers across 28 providers. Compare model IDs, prices and limits in Providers.

What is the context window?

The cataloged model context is 262,144 tokens. Each serving endpoint may apply a different limit.

Does it support tools and structured output?

Tool calling: Not reported. Structured output: Not reported. Support can vary by endpoint.

Resources

Hugging Face ↗
Catalog history · 3 recorded revisions

Metadata observations since this model was first cataloged. These are not model release versions.

  • · 20e9f5decee6
  • · b5ad30ea35c8
  • · e61c27faf2b5