Gemma 4 26B A4B IT

Released
2026-03-31
Last refreshed
2026-06-30
Status
Researched 164d ago
Open sourceCommercial use: permittedMultimodalCodingRAGAgentsLong contextVisionClassificationJSON / Tool use
Google DeepMind releases · 54 in the last 12 months · this family litChangelog →
Specifications
Family
Gemma 4
Released
2026-03-31
Context
256K
Parameters
26B
Knowledge cutoff
2025-01
Specialization
general
Openness
Open source
License
Apache 2.0OSI-approvedPermitted
Weights
Available
Code
Unknown
Created by

Pioneering artificial intelligence research.

London, United Kingdom
Founded 2014
Website
Pricing
Output / 1M
Free
Input / 1M
Free

Cheapest of 9 routes · Open Inference

About

Instruction-tuned MoE variant with 26B parameters (4B active). Optimized for instruction-following with advanced reasoning and multimodal capabilities.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ A

Providers(9)

Compare all 9

Compare API pricing across 9 tracked provider routes for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MCacheRoute
Open InferenceFreeFree-
Serverless
NextBit$0.090$0.300read $0.050
Serverless
Cloudflare Workers AI$0.100$0.300-
Serverless
Novita AI$0.130$0.400-
Serverless
Vercel AI Gateway$0.130$0.400-
Serverless
GCP Vertex AI$0.150$0.600-
Serverless
OpenRouter$0.150$0.600-
Serverless
Google AI Studio---
ServerlessPartial
Together AI---
ServerlessPartial

Available via routers & gateways(13)

LiteLLM

Gateway

Open-source Python SDK and proxy server that unifies 100+ LLM APIs behind a single OpenAI-compatible interface, with load balancing, cost tracking, and configurable failover.

Free OSSGoogle AI StudioGCP Vertex AI

OpenRouter

Hybrid

Unified hybrid gateway to 500+ models from 80+ providers via a single OpenAI-compatible API, with an optional Auto Router that selects a model per prompt from market spend data.

Passthrough + feeGoogle AI StudioGCP Vertex AITogether AI

Portkey

Gateway

Production AI gateway routing to 1,600+ LLMs with failover, load balancing, caching, and guardrails; MIT-licensed open-source core is self-hostable. Now part of Palo Alto Networks as Prisma AIRS AI Gateway.

SubscriptionGoogle AI StudioGCP Vertex AI

AIRouter

Router

Commercial LLM router (branded "AI Router" on its site) that picks the best model per request via a Decision API or an OpenAI-compatible Chat Completions endpoint, with a private selection mode that routes on anonymized patterns; sold as BYOK subscription plans priced by routed tokens.

SubscriptionGoogle AI StudioGCP Vertex AI

Helicone

Gateway

Observability-first AI gateway with routing, caching, rate limiting, and request tracing; Apache-2.0 open-source platform with a managed hosted tier for logging and analytics.

SubscriptionGoogle AI StudioGCP Vertex AI

Kong AI Gateway

Gateway

Multi-LLM AI gateway built on Kong Gateway 3.x, adding semantic routing, load balancing, guardrails, and MCP traffic analytics as plugins over Kong's existing API management platform.

SubscriptionGoogle AI StudioGCP Vertex AI

Capabilities

VisionMultimodalJSON / Tool useStructured Outputs

Benchmark peer barsfor Coding

Benchmark scores(9)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
88.3AIME 2026 (accuracy)Observed 2026-06-07—
77.1LiveCodeBench v6 (accuracy)Observed 2026-06-07—
82.6From official Gemma 4 model card (accuracy)Observed 2026-06-07—
1438.0text_style_control overall 2026-10-22 (style control, default view)Observed 2026-10-26
Configuration: Gemma 4 26B A4B
Harness: LMArena blind pairwise human votes, Bradley-Terry
Evaluator: LMArena
Confidence: confirmed
Notes: LMArena text style control leaderboard, overall category, published 2026-10-22 (HF lmarena-ai/leaderboard-dataset @9761a3ce, latest split). Board entry 'Gemma 4 26B A4B': rank 100, Arena Score 1437.5 (95% CI 1430.1-1444.9), 6092 votes. Retrieved 2026-10-26.
Source

Source: LMArena (lmarena.ai), CC BY 4.0

1242.0vision_style_control overall 2026-10-19 (style control, default view)Observed 2026-10-26
Configuration: Gemma 4 26B A4B
Harness: LMArena blind pairwise human votes, Bradley-Terry
Evaluator: LMArena
Confidence: confirmed
Notes: LMArena vision style control leaderboard, overall category, published 2026-10-19 (HF lmarena-ai/leaderboard-dataset @9761a3ce, latest split). Board entry 'Gemma 4 26B A4B': rank 65, Arena Score 1241.7 (95% CI 1235.3-1248.2), 23922 votes. Retrieved 2026-10-26.
Source

Source: LMArena (lmarena.ai), CC BY 4.0

1401.0text creative_writing 2026-10-22Observed 2026-10-22
Configuration: Gemma 4 26B A4B
Harness: LMArena blind pairwise human votes, Bradley-Terry
Evaluator: LMArena
Confidence: confirmed
Notes: LMArena Text arena, Creative Writing category (no style control), board published 2026-10-22 (HF lmarena-ai/leaderboard-dataset @a467e6b9, text config, latest split). Board entry 'Gemma 4 26B A4B' (category creative_writing in text/latest-00000-of-00001.parquet): rank 97 of 416, rating 1401.2 (95% CI 1383.0-1419.3), 1000 votes. Source: LMArena (CC BY 4.0). Retrieved 2026-10-27.
Source

Source: LMArena (lmarena.ai), CC BY 4.0

82.3Diamond (self-reported, official Gemma 4 model card)Observed 2026-11-05
Confidence: self-reported
Notes: Self-reported by Google on the official Gemma 4 model card (Benchmark Results table, column 'Gemma 4 26B A4B'); the card states results are for instruction-tuned models. Same table on https://huggingface.co/google/gemma-4-26b-a4b-it README and https://ai.google.dev/gemma/docs/core/model_card_4. Retrieved 2026-11-05. GPQA Diamond 82.3%.
8.7HLE no tools (self-reported, official Gemma 4 model card)Observed 2026-11-05
Confidence: self-reported
Notes: Self-reported by Google on the official Gemma 4 model card (Benchmark Results table, column 'Gemma 4 26B A4B'); the card states results are for instruction-tuned models. Same table on https://huggingface.co/google/gemma-4-26b-a4b-it README and https://ai.google.dev/gemma/docs/core/model_card_4. Retrieved 2026-11-05. HLE no tools 8.7%.
73.8MMMU Pro (self-reported, official Gemma 4 model card)Observed 2026-11-05
Confidence: self-reported
Notes: Self-reported by Google on the official Gemma 4 model card (Benchmark Results table, column 'Gemma 4 26B A4B'); the card states results are for instruction-tuned models. Same table on https://huggingface.co/google/gemma-4-26b-a4b-it README and https://ai.google.dev/gemma/docs/core/model_card_4. Retrieved 2026-11-05. MMMU Pro 73.8% (Vision section; no Vision-only setting stated).

Migration checks

No linked migration route is available for this model yet.

Compare Gemma 4 26B A4B IT with other models