---
title: "Cheaper alternatives to o3 · Undominated.ai"
canonical: https://undominated.ai/alternatives/openai__o3/
description: "51 models score at least as high as o3 and cost no more on a balanced workload. 7 give up nothing measurable; the other 44 name what they drop. Measured prices and independent scores."
---

# Cheaper alternatives to o3 · Undominated.ai

> 51 models score at least as high as o3 and cost no more on a balanced workload. 7 give up nothing measurable; the other 44 name what they drop. Measured prices and independent scores.

# Cheaper alternatives to o3

OpenAI · LMArena 1409.2 · $3.50/M on a balanced workload · prices as of 2026-08-28

## 7 models are 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, **51** score at least as high as o3 and cost no more under at least one workload. **7** match or beat it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. **44** 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 o3, it is not generated.

## Where it holds

| Workload | o3 $/M | Better and cheaper | With a trade |
| --- | --- | --- | --- |
| Balanced | $3.50 | 7 | 44 |
| Summarise | $1.87 | 7 | 44 |
| Chat | $3.95 | 6 | 42 |
| Code gen | $5.48 | 6 | 42 |
| Agentic | $2.01 | 6 | 44 |

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 | 1409.2 |
| --- | --- |
| Context window | 200K |
| Max output | 100K |
| Input modes | image, text, file |
| Tool use | yes |
| Extended reasoning | yes |

A replacement has to clear every line above, not just the score. [Full record for o3](/models/openai__o3/).

## Better and cheaper, nothing given up 7

Each of these matches or beats o3 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 | Muse Spark 1.1 Meta | 1478.3 +69.1 | $2.00 | Balanced −43% Summarise −42% Chat −46% Code gen −46% Agentic −48% |
| 2 | Grok 4.5 xAI | 1452.3 +43.1 | $3.00 | Balanced −14% Summarise −8% Chat −22% Code gen −22% Agentic −21% |
| 3 | Grok 4.6 xAI | 1443.7 +34.5 | $3.00 | Balanced −14% Summarise −5% Chat −20% Code gen −22% Agentic −15% |
| 4 | GPT-5.1 OpenAI | 1441.4 +32.2 | $3.44 | Balanced −2% Summarise −27% Agentic −6% |
| 5 | GPT-5.6 Luna OpenAI | 1428.5 +19.3 | $0.450 | Balanced −87% Summarise −89% Chat −86% Code gen −86% Agentic −88% |
| 6 | Mistral Medium 3.5 Mistral | 1420.9 +11.7 | $3.00 | Balanced −14% Summarise −4% Chat −1% Code gen −7% |
| 7 | GPT-5.4 Mini OpenAI | 1412.1 tie | $1.69 | Balanced −52% Summarise −60% Chat −48% Code gen −46% Agentic −55% |

## Cheaper and higher-scoring, but you give something up 44

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 |
| --- | --- | --- | --- | --- |
| Gemini 3.7 Flash Google | 1490.2 +81 | $0.750 | Balanced −79% Summarise −81% Chat −78% Code gen −77% Agentic −80% | 100K → 66K max output |
| Qwen3.8 Max Qwen | 1481.9 +72.7 | $3.00 | Balanced −14% Summarise −9% Chat −22% Code gen −22% Agentic −22% | no file input |
| Gemini 3.5 Flash Google | 1482.6 +73.4 | $3.38 | Balanced −4% Summarise −20% Agentic −9% | 100K → 66K max output |
| Gemini 3.6 Flash Google | 1476.5 +67.3 | $1.50 | Balanced −57% Summarise −62% Chat −56% Code gen −54% Agentic −60% | 100K → 66K max output |
| GLM 5.3 Z.ai | 1476.9 +67.7 | $2.15 | Balanced −39% Summarise −35% Chat −43% Code gen −43% Agentic −42% | no image, file input |
| MiMo-V2.5-Pro Xiaomi | 1465 +55.8 | $0.544 | Balanced −84% Summarise −82% Chat −88% Code gen −88% Agentic −88% | no image, file input |
| GLM 5.2 Z.ai | 1465.4 +56.2 | $1.83 | Balanced −48% Summarise −44% Chat −51% Code gen −52% Agentic −50% | no image, file input |
| GLM 5.1 Z.ai | 1464.1 +54.9 | $1.94 | Balanced −45% Summarise −41% Chat −49% Code gen −49% Agentic −47% | no image, file input |
| Qwen3.7 Plus Qwen | 1456.2 +47 | $0.560 | Balanced −84% Summarise −84% Chat −84% Code gen −84% Agentic −84% | no file input |
| Kimi K2.6 Moonshot AI | 1455 +45.8 | $1.71 | Balanced −51% Summarise −53% Chat −51% Code gen −50% Agentic −53% | no file input |
| Gemini 2.5 Pro Google | 1457.3 +48.1 | $3.44 | Balanced −2% Summarise −27% Agentic −6% | 100K → 66K max output |
| Kimi K2.5 Moonshot AI | 1445.2 +36 | $0.900 | Balanced −74% Summarise −77% Chat −73% Code gen −73% Agentic −75% | no file input |
| GLM 5 Z.ai | 1445.2 +36 | $0.930 | Balanced −73% Summarise −72% Chat −75% Code gen −75% Agentic −74% | no image, file input |
| Gemma 4 31B Google | 1441.7 +32.5 | $0.152 | Balanced −96% Summarise −95% Chat −95% Code gen −96% Agentic −95% | 100K → 16K max output no file input |
| Hy3 Tencent | 1441.2 +32 | $0.144 | Balanced −96% Summarise −96% Chat −96% Code gen −96% Agentic −96% | no image, file input |
| Qwen3.6 Max Preview Qwen | 1446.3 +37.1 | $2.31 | Balanced −34% Summarise −31% Chat −22% Code gen −25% Agentic −10% | 100K → 66K max output no image, file input |
| Qwen3.8 27B Qwen | 1440.8 +31.6 | $0.956 | Balanced −73% Summarise −77% Chat −70% Code gen −69% Agentic −73% | no file input |
| GLM 4.6 Z.ai | 1439.8 +30.6 | $0.875 | Balanced −75% Summarise −75% Chat −75% Code gen −75% Agentic −76% | no image, file input |
| DeepSeek V4 Pro 0423 DeepSeek | 1439.2 +30 | $1.09 | Balanced −69% Summarise −63% Chat −75% Code gen −76% Agentic −74% | no image, file input |
| Qwen3.5 397B A17B Qwen | 1438.3 +29.1 | $0.877 | Balanced −75% Summarise −74% Chat −70% Code gen −72% Agentic −66% | 100K → 66K max output no file input |
| Qwen3.6 Plus Qwen | 1436.8 +27.6 | $0.731 | Balanced −79% Summarise −78% Chat −75% Code gen −76% Agentic −72% | 100K → 66K max output no file input |
| Inkling Thinking Machines | 1439.2 +30 | $1.73 | Balanced −51% Summarise −53% Chat −51% Code gen −50% Agentic −53% | no file input |
| Gemma 4 26B A4B Google | 1434.6 +25.4 | $0.138 | Balanced −96% Summarise −96% Chat −95% Code gen −96% Agentic −94% | 100K → 16K max output no file input |
| Gemini 3.5 Flash Lite Google | 1436.5 +27.3 | $0.850 | Balanced −76% Summarise −82% Chat −72% Code gen −71% Agentic −77% | 100K → 66K max output |
| MiniMax M3 MiniMax | 1434.8 +25.6 | $0.525 | Balanced −85% Summarise −85% Chat −85% Code gen −85% Agentic −85% | no file input |
| GLM 4.7 Z.ai | 1435.3 +26.1 | $0.738 | Balanced −79% Summarise −80% Chat −79% Code gen −78% Agentic −79% | no image, file input |
| GLM 5V Turbo Z.ai | 1436.8 +27.6 | $1.90 | Balanced −46% Summarise −43% Chat −49% Code gen −49% Agentic −48% | no file input |
| DeepSeek V4 Flash 0423 DeepSeek | 1431.6 +22.4 | $0.101 | Balanced −97% Summarise −96% Chat −98% Code gen −98% Agentic −97% | no image, file input |
| MiMo-V2.5 Xiaomi | 1427.3 +18.1 | $0.175 | Balanced −95% Summarise −94% Chat −96% Code gen −96% Agentic −96% | no file input |
| GLM 4.5 Z.ai | 1429.4 +20.2 | $1.00 | Balanced −71% Summarise −71% Chat −72% Code gen −72% Agentic −73% | 200K → 131K context 100K → 98K max output no image, file input |
| R1 0528 DeepSeek | 1427.9 +18.7 | $0.912 | Balanced −74% Summarise −71% Chat −72% Code gen −73% Agentic −67% | 200K → 164K context 100K → 33K max output no image, file input |
| DeepSeek V3.2 DeepSeek | 1424.6 +15.4 | $0.290 | Balanced −92% Summarise −88% Chat −93% Code gen −94% Agentic −90% | 200K → 164K context no image, file input |
| DeepSeek V3.2 Exp DeepSeek | 1424.4 +15.2 | $0.305 | Balanced −91% Summarise −85% Chat −92% Code gen −94% Agentic −86% | 200K → 164K context 100K → 66K max output no image, file input |
| Qwen3 VL 235B A22B Instruct Qwen | 1420.9 +11.7 | $0.632 | Balanced −82% Summarise −86% Chat −78% Code gen −78% Agentic −80% | 100K → 33K max output no file input no extended reasoning |
| Qwen3 235B A22B Instruct 2507 Qwen | 1419.3 tie | $0.205 | Balanced −94% Summarise −94% Chat −93% Code gen −93% Agentic −92% | 100K → 16K max output no image, file input no extended reasoning |
| DeepSeek V3.1 Terminus DeepSeek | 1419.6 tie | $0.453 | Balanced −87% Summarise −86% Chat −87% Code gen −87% Agentic −85% | 200K → 164K context 100K → 33K max output no image, file input |
| Qwen3 Next 80B A3B Instruct Qwen | 1418.6 tie | $0.350 | Balanced −90% Summarise −92% Chat −88% Code gen −87% Agentic −88% | no image, file input no extended reasoning |
| DeepSeek V3.1 DeepSeek | 1419.1 tie | $0.825 | Balanced −76% Summarise −68% Chat −75% Code gen −78% Agentic −64% | 200K → 164K context no image, file input |
| Qwen3.5-122B-A10B Qwen | 1417.9 tie | $0.715 | Balanced −80% Summarise −81% Chat −75% Code gen −75% Agentic −73% | no file input |
| Gemini 2.5 Flash Google | 1417.3 tie | $0.850 | Balanced −76% Summarise −82% Chat −72% Code gen −71% Agentic −77% | 100K → 66K max output |
| Gemini 3.1 Flash Lite Preview Google | 1414.8 tie | $0.563 | Balanced −84% Summarise −87% Chat −83% Code gen −82% Agentic −85% | 100K → 66K max output |
| Qwen3 235B A22B Thinking 2507 Qwen | 1413.8 tie | $0.748 | Balanced −79% Summarise −82% Chat −73% Code gen −73% Agentic −73% | 200K → 131K context no image, file input |
| Inkling Small Thinking Machines | 1411.7 tie | $0.637 | Balanced −82% Summarise −79% Chat −84% Code gen −84% Agentic −82% | no file input |
| Qwen3 Max Qwen | 1412.7 tie | $1.56 | Balanced −55% Summarise −60% Chat −53% Code gen −53% Agentic −56% | 100K → 66K max output no image, file 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/).