---
title: "Cheaper alternatives to Gemma 3 27B · Undominated.ai"
canonical: https://undominated.ai/alternatives/google__gemma-3-27b-it/
description: "10 models score at least as high as Gemma 3 27B and cost no more on a balanced workload. 1 give up nothing measurable; the other 9 name what they drop. Measured prices and independent scores."
---

# Cheaper alternatives to Gemma 3 27B · Undominated.ai

> 10 models score at least as high as Gemma 3 27B and cost no more on a balanced workload. 1 give up nothing measurable; the other 9 name what they drop. Measured prices and independent scores.

# Cheaper alternatives to Gemma 3 27B

Google · LMArena 1358.3 · $0.172/M on a balanced workload · prices as of 2026-08-28

## 1 model is both better and cheaper, giving up nothing.

Of the 133 models carrying both an independent score and a published price at the same delivery mode, **10** score at least as high as Gemma 3 27B and cost no more under at least one workload. **1** matches or beats it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. **9** do not, and every row names what it drops.

Dominance is not a property of a model. It is a property of a model, a capability lens and a workload, and all three are named on every row. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset; prices are blended per million tokens. A higher score is not a drop-in replacement.

This page is built from that comparison and nothing else. When no model is both better and cheaper than Gemma 3 27B, it is not generated.

## Where it holds

| Workload | Gemma 3 27B $/M | Better and cheaper | With a trade |
| --- | --- | --- | --- |
| Balanced | $0.172 | 0 | 9 |
| Summarise | $0.087 | 0 | 7 |
| Chat | $0.216 | 1 | 9 |
| Code gen | $0.299 | 1 | 9 |
| Agentic | $0.112 | 1 | 8 |

Workload mixes are defined on the [methodology page](/methodology/). Cached input is priced at the cached rate, and reasoning tokens at the worse of the reasoning and output rates.

## What you would be replacing

| Intelligence | 1358.3 |
| --- | --- |
| Context window | 262K |
| Max output | 118K |
| Input modes | text, image |
| Tool use | yes |
| Extended reasoning | no |

A replacement has to clear every line above, not just the score. [Full record for Gemma 3 27B](/models/google__gemma-3-27b-it/).

## Better and cheaper, nothing given up 1

Each of these matches or beats Gemma 3 27B on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

| # | Model | Intelligence | $/M | Holds under |
| --- | --- | --- | --- | --- |
| 1 | MiMo-V2.5 Xiaomi | 1427.3 +69 | $0.155 chat | Chat −28% Code gen −29% Agentic −29% |

## Cheaper and higher-scoring, but you give something up 9

These score at least as high and cost no more on the two plotted axes, and lose something that is not on them. Read the last column before switching.

| Model | Intelligence | $/M | Holds under | What you give up |
| --- | --- | --- | --- | --- |
| Gemma 4 31B Google | 1441.7 +83.4 | $0.152 | Balanced −12% Chat −18% Code gen −21% Agentic −7% | 118K → 16K max output |
| Hy3 Tencent | 1441.2 +82.9 | $0.144 | Balanced −16% Summarise −11% Chat −25% Code gen −24% Agentic −26% | no image input |
| Gemma 4 26B A4B Google | 1434.6 +76.3 | $0.138 | Balanced −20% Summarise −4% Chat −18% Code gen −22% Agentic −1% | 118K → 16K max output |
| DeepSeek V4 Flash 0423 DeepSeek | 1431.6 +73.3 | $0.101 | Balanced −41% Summarise −24% Chat −56% Code gen −58% Agentic −51% | no image input |
| Step 3.5 Flash StepFun | 1403.8 +45.5 | $0.150 | Balanced −13% Chat −17% Code gen −26% | 118K → 66K max output no image input |
| Qwen3.5-Flash Qwen | 1397.6 +39.3 | $0.114 | Balanced −34% Summarise −14% Chat −34% Code gen −39% Agentic −16% | 118K → 66K max output |
| Qwen3 30B A3B Instruct 2507 Qwen | 1384.3 +26 | $0.084 | Balanced −51% Summarise −36% Chat −51% Code gen −55% Agentic −37% | 118K → 32K max output no image input |
| Solar Pro 4 Upstage | 1376.2 +17.9 | $0.052 | Balanced −70% Summarise −68% Chat −73% Code gen −73% Agentic −74% | no image input |
| gpt-oss-120b OpenAI | 1365.6 tie | $0.070 | Balanced −59% Summarise −50% Chat −58% Code gen −61% Agentic −49% | 262K → 131K context no image input |

## What this compares, and what it leaves out

 - Quality is LMArena Elo, used under CC BY 4.0 from the official dataset. The 95% confidence interval on a difference between two scores is about ±10.68 points, so a gap smaller than that is marked *tie* rather than an improvement — see [significance bands](/significance/).
 - 188 further models at this delivery mode carry a price but no independent score. They are absent from the comparison in both directions — unrated is not a zero, and an unmeasured model is neither an alternative nor a worse buy.
 - Retired models are never offered as an alternative, and a model only competes against its own delivery mode: batch trades latency for price, so it is not a like-for-like swap.
 - Nothing here measures latency, throughput, rate limits or how a model behaves on your prompts. Two models with the same index score are not interchangeable.
 - The models that nothing beats on both axes are on the [value frontier](/frontier/), and every other model something cheaper beats is [listed here](/alternatives/).