---
title: "Yeltsin in the AI Aisle"
description: "OpenAI's year-old GPT-OSS commands 36% of Opus 4.8's OpenRouter volume. The reason is segmentation: buyers split across cost, speed \u0026 accuracy, \u0026 mixture-of-experts just moved the local boundary."
categories: ["AI","local models","open source"]
keywords: ["GPT-OSS-120b","GLM 5.2","Claude Opus 4.8","OpenRouter","mixture of experts","inference market segmentation","local LLM","Laguna S 2.1"]
ai_summary: "A year-old GPT-OSS-120b still serves 36% of Claude Opus 4.8's daily token volume on OpenRouter because the inference market has segmented across cost, speed, \u0026 accuracy. GLM 5.2 now out-serves the frontier at 495B tokens a day, \u0026 Opus 5 shipped to contest the open-weight field. Mixture-of-experts lets a 118B model decode at the cost of a 26B model, moving the boundary of the local segment."
date: 2026-07-23
lastmod: 2026-07-27
canonical_url: https://www.tomtunguz.com/yeltsin-in-the-ai-aisle/
author: "Tomasz Tunguz"
---


In 1989 Boris Yeltsin stopped at a Randalls supermarket in Houston, stunned by the variety of ice cream.[^1] OpenRouter is that supermarket aisle for AI.

And shoppers make surprising choices : OpenAI's year-old open-source model GPT-OSS 120b[^2] commands 36% of Anthropic's Opus 4.8 volume.[^3]

Why does a model from August 2025 still hold a third of the traffic of a frontier model that shipped weeks ago?

The market for tokens has segmented.

{{< email_image src="ft8lhobzy5krc1fmwqyl" alt="Daily tokens served on OpenRouter, 7-day average: GLM 5.2 tops the field at 495B, above Claude Opus 4.8 at 199.6B; a year-old GPT-OSS-120b holds 71.3B, roughly a third of Opus, with Gemma 4 26B at 50B & the rest of the open field trailing to Phi-4 at 90M" width="540" height="375" >}}

Segmentation happens because buyer needs vary.

The segmentation is accelerating driven by competition. Last week Anthropic shipped Opus 5, smaller & cheaper than Fable,[^4] explicitly to contest the ground that Moonshot's Kimi 3 targets.[^5] In the mid-model-market, Poolside launched Laguna S 2.1, a US mid-market model.[^6]

Size (small, medium, large, XL), origin (US v China), architecture (dense vs sparse), accuracy (coding focused or general), speed (tokens per second), modality (text-only or vision) ; there are many flavors of AI.

I spent the weekend replacing the model that runs my agent. The incumbent is Gemma 4 26b; the challenger is Laguna S 2.1, a 118-billion-parameter model. By every number I expected to matter, the 118b model should have been slower.

On my M5 Max both generate at the same speed, because Laguna is a mixture-of-experts architecture: 118 billion parameters live in memory, but only 8 billion activate per token. A 118b model now runs at the decode cost of a 26b model, which pulls frontier-class quality down into the local tier.

The accuracy shows up where it matters. My local stack runs a coding & email agent on tool calls, & across the models I have cycled through, the tool-call failure rate falls from 29.4% to 20.1% as active parameters climb.[^7] Laguna reduces error rates by 7 percentage points over the 26b model it replaced.

{{< email_image src="swrsh8zytby9k3shublz" alt="MCP tool-call failure rate for the local model driving a production automation stack, measured from real traffic: Ornith-1.0 35B at 29.4%, Gemma 4 26B at 27.1%, & Laguna S 2.1 at 20.1%, roughly a quarter fewer failures than Gemma" width="540" height="318" >}}

Segmentation is the sign of a healthy competitive market. The frontier still serves the world's hardest tokens. It no longer has to serve all of them, & the tier on my laptop just got a much higher ceiling ; a trend that competition will push forward inexorably.

[^1]: [Boris Yeltsin's 1989 visit to a Houston grocery store](https://www.houstonpublicmedia.org/articles/shows/houston-matters/2020/02/21/361467/boris-yelstins-1989-visit-to-a-houston-grocery-store-is-now-an-opera/), a Randalls in Clear Lake, September 16, 1989.

[^2]: [Introducing gpt-oss (OpenAI)](https://openai.com/index/introducing-gpt-oss/), released August 5, 2025.

[^3]: OpenRouter model activity pages, 7-day average, retrieved 2026-07-27: [GPT-OSS-120b](https://openrouter.ai/openai/gpt-oss-120b/activity), [GLM 5.2](https://openrouter.ai/z-ai/glm-5.2/activity), [Claude Opus 4.8](https://openrouter.ai/anthropic/claude-opus-4.8/activity).

[^4]: [Anthropic debuts Claude Opus 5 at half the price](https://www.technology.org/2026/07/27/anthropic-claude-opus-5-launch-half-price/), launched July 24, 2026.

[^5]: [Moonshot's Kimi 3 is expected to close the gap with Anthropic's Opus 4.8 (TechCrunch)](https://techcrunch.com/2026/07/16/moonshots-upcoming-kimi-3-is-expected-to-close-the-gap-with-anthropics-opus-4-8/).

[^6]: [Introducing Laguna S 2.1 (Poolside)](https://poolside.ai/blog/introducing-laguna-s-2-1), a 118B-total, 8B-active open-weight model released July 22, 2026.

[^7]: Author's production data: MCP tool-call logs from a local coding & email agent, measured across three local models over five months.