---
title: "Intelligence Per Dollar"
description: "Microsoft's MAI-Code-1-Flash matches Claude Haiku 4.5 on SWE-Bench Verified using a third of the tokens. The right metric for AI in production is intelligence per dollar, value per token spent."
categories: ["AI","metrics"]
keywords: ["tokens per result","token efficiency AI","Microsoft MAI-Code-1-Flash","token cost optimization","inference cost per task","token-per-result metric","AI coding models","token efficiency benchmarks","SWE-Bench token efficiency"]
ai_summary: "Microsoft's MAI-Code-1-Flash achieves 60% fewer tokens than Claude Haiku 4.5 on hard coding tasks (SWE-Bench Verified), a +16-point lead on SWE-Bench Pro, and adaptive solution length control. The key insight: intelligence per dollar, value per token spent, is the right metric for evaluating AI models in production."
date: 2026-06-03
lastmod: 2026-07-17
canonical_url: https://www.tomtunguz.com/tokens-per-result/
author: "Tomasz Tunguz"
---


{{< email_image src="fczixgwvrhqhqt5uxfto" alt="Screenshot 2026-06-02 at 9.22.43 PM" width="540" height="288" >}}

Yesterday Microsoft added a new metric to a model release card, one that will likely become a standard.[^7]

Average token usage.

In the first row, the Microsoft model hits 71.6 on SWE-Bench Verified using about a third of the tokens Claude Haiku 4.5 burns.

Benchmarks are now measured on two different dimensions, the overall performance & the cost to achieve that intelligence.

This is yet another sign that the era of subsidies[^1], tokenmaxxing[^2], & all-out performance for many use cases is over.

Even the most valuable companies in the world cannot afford state-of-the-art intelligence for every conceivable use case.[^5] Uber capped employee AI spending after blowing through its budget in four months.[^3] Salesforce is spending $300M on Anthropic tokens & has frozen engineering hires.[^4]

This new dual benchmark answers the buyer's only question : what is my intelligence per dollar?

{{< email_image src="hnlfpw6c8qaurohqluul" alt="Screenshot 2026-06-03 at 5.49.00 AM" width="540" height="270" >}}

Artificial Analysis already benchmarks this.[^6] GPT 5.5 & Claude Opus 4.8 land within a point of each other on the Intelligence Index, around 60. Running the index costs $3,357 on GPT 5.5 & $4,685 on Opus 4.8. Same answer, 40% more expensive.

Model companies must now compete on both dimensions. The application layer will compete one level up, on dollars per outcome, what a closed ticket, a shipped PR, or a resolved support case actually costs.

Every layer in the stack now has to price the same way the customer thinks : per result, not per token.

---

[^1]: [The Unsustainable Subsidy](https://tomtunguz.com/ai-model-inflation/) — The era of AI subsidies is ending.
[^2]: [Tokenmaxxing](https://tomtunguz.com/tokenmaxxing/) — Models that game benchmarks with extra tokens are losing their edge.
[^3]: [Uber caps employee AI spending after blowing through budget in 4 months](https://techcrunch.com/2026/06/02/uber-caps-employee-ai-spending-after-blowing-through-budget-in-four-months/) — Uber caps employee AI spending after blowing through budget in four months.
[^4]: [Salesforce Spends $300M on AI, Freezes Engineering Hires](https://enterprisedna.co/resources/news/salesforce-300m-anthropic-tokens-engineer-hiring-freeze-2026/) — Salesforce Spends $300M on AI, Freezes Engineering Hires.
[^5]: [Microsoft cancels Claude Code licenses, shifting developers to GitHub Copilot CLI](https://www.windowscentral.com/microsoft/microsoft-cancels-claude-code-licenses-shifting-developers-to-github-copilot-cli-a-move-likely-driven-by-financial-motives) — Microsoft cancelled Claude Code licenses across its Experiences and Devices division (Windows, Microsoft 365, Outlook, Teams, Surface) after engineering usage outran budgets.

[^6]: [AI Model & API Providers Analysis](https://artificialanalysis.ai/) — Independent analysis of AI model costs.

[^7]: [Introducing MAI-Code-1-Flash](https://microsoft.ai/news/introducingmai-code-1-flash/) — Microsoft announces a new coding model with average token usage on the release card.