---
title: "Cheaper alternatives to Inkling Small · Undominated.ai"
canonical: https://undominated.ai/alternatives/thinkingmachines__inkling-small/
description: "22 models score at least as high as Inkling Small and cost no more on a balanced workload. 0 give up nothing measurable; the other 22 name what they drop. Measured prices and independent scores."
---

# Cheaper alternatives to Inkling Small · Undominated.ai

> 22 models score at least as high as Inkling Small and cost no more on a balanced workload. 0 give up nothing measurable; the other 22 name what they drop. Measured prices and independent scores.

# Cheaper alternatives to Inkling Small

Thinking Machines · LMArena 1411.7 · $0.637/M on a balanced workload · prices as of 2026-08-28

## Everything cheaper and higher-scoring gives something up.

Of the 133 models carrying both an independent score and a published price at the same delivery mode, **22** score at least as high as Inkling Small and cost no more under at least one workload. **22** 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 Inkling Small, it is not generated.

## Where it holds

| Workload | Inkling Small $/M | Better and cheaper | With a trade |
| --- | --- | --- | --- |
| Balanced | $0.637 | 0 | 16 |
| Summarise | $0.388 | 0 | 22 |
| Chat | $0.645 | 0 | 14 |
| Code gen | $0.872 | 0 | 13 |
| Agentic | $0.354 | 0 | 15 |

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 | 1411.7 |
| --- | --- |
| Context window | 1.0M |
| Max output | 262K |
| Input modes | text, image, audio |
| Tool use | yes |
| Extended reasoning | yes |

A replacement has to clear every line above, not just the score. [Full record for Inkling Small](/models/thinkingmachines__inkling-small/).

## Cheaper and higher-scoring, but you give something up 22

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 |
| --- | --- | --- | --- | --- |
| MiMo-V2.5-Pro Xiaomi | 1465 +53.3 | $0.544 | Balanced −15% Summarise −14% Chat −26% Code gen −24% Agentic −31% | 262K → 131K max output no image, audio input |
| Qwen3.7 Plus Qwen | 1456.2 +44.5 | $0.560 | Balanced −12% Summarise −24% Chat −3% Agentic −12% | 1.0M → 1.0M context 262K → 131K max output no audio input |
| Gemma 4 31B Google | 1441.7 +30 | $0.152 | Balanced −76% Summarise −77% Chat −72% Code gen −73% Agentic −71% | 1.0M → 262K context 262K → 16K max output no audio input |
| Hy3 Tencent | 1441.2 +29.5 | $0.144 | Balanced −77% Summarise −80% Chat −75% Code gen −74% Agentic −77% | 1.0M → 262K context 262K → 128K max output no image, audio input |
| Gemma 4 26B A4B Google | 1434.6 +22.9 | $0.138 | Balanced −78% Summarise −78% Chat −72% Code gen −73% Agentic −69% | 1.0M → 262K context 262K → 16K max output no audio input |
| DeepSeek V4 Flash 0423 DeepSeek | 1431.6 +19.9 | $0.101 | Balanced −84% Summarise −83% Chat −85% Code gen −86% Agentic −85% | no image, audio input |
| MiniMax M3 MiniMax | 1434.8 +23.1 | $0.525 | Balanced −18% Summarise −29% Chat −9% Code gen −6% Agentic −18% | no audio input |
| MiMo-V2.5 Xiaomi | 1427.3 +15.6 | $0.175 | Balanced −73% Summarise −72% Chat −76% Code gen −76% Agentic −78% | 262K → 131K max output |
| GPT-5.6 Luna OpenAI | 1428.5 +16.8 | $0.450 | Balanced −29% Summarise −49% Chat −15% Code gen −10% Agentic −31% | 262K → 128K max output no audio input |
| DeepSeek V3.2 DeepSeek | 1424.6 +12.9 | $0.290 | Balanced −55% Summarise −41% Chat −58% Code gen −63% Agentic −43% | 1.0M → 164K context 262K → 147K max output no image, audio input |
| DeepSeek V3.2 Exp DeepSeek | 1424.4 +12.7 | $0.305 | Balanced −52% Summarise −29% Chat −49% Code gen −59% Agentic −18% | 1.0M → 164K context 262K → 66K max output no image, audio input |
| Qwen3 235B A22B Instruct 2507 Qwen | 1419.3 tie | $0.205 | Balanced −68% Summarise −71% Chat −58% Code gen −58% Agentic −55% | 1.0M → 262K context 262K → 16K max output no image, audio input no extended reasoning |
| Qwen3 Next 80B A3B Instruct Qwen | 1418.6 tie | $0.350 | Balanced −45% Summarise −64% Chat −24% Code gen −20% Agentic −34% | 1.0M → 262K context 262K → 236K max output no image, audio input no extended reasoning |
| DeepSeek V3.1 Terminus DeepSeek | 1419.6 tie | $0.453 | Balanced −29% Summarise −31% Chat −19% Code gen −20% Agentic −16% | 1.0M → 164K context 262K → 33K max output no image, audio input |
| Qwen3 VL 235B A22B Instruct Qwen | 1420.9 tie | $0.632 | Balanced −1% Summarise −32% | 1.0M → 262K context 262K → 33K max output no audio input no extended reasoning |
| Gemini 3.1 Flash Lite Preview Google | 1414.8 tie | $0.563 | Balanced −12% Summarise −36% Agentic −14% | 262K → 66K max output |
| Gemini 3.7 Flash Google | 1490.2 +78.5 | $0.354 summarise | Summarise −9% | 262K → 66K max output |
| Gemini 3.5 Flash Lite Google | 1436.5 +24.8 | $0.333 summarise | Summarise −14% | 262K → 66K max output |
| GLM 4.7 Z.ai | 1435.3 +23.6 | $0.376 summarise | Summarise −3% | 1.0M → 205K context 262K → 131K max output no image, audio input |
| Qwen3.5-122B-A10B Qwen | 1417.9 tie | $0.351 summarise | Summarise −9% | 1.0M → 262K context 262K → 236K max output no audio input |
| Gemini 2.5 Flash Google | 1417.3 tie | $0.333 summarise | Summarise −14% | 262K → 66K max output |
| Qwen3 235B A22B Thinking 2507 Qwen | 1413.8 tie | $0.334 summarise | Summarise −14% | 1.0M → 131K context 262K → 118K max output no image, audio 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/).