news

Share Of US Models Being Used On OpenRouter Has Collapsed From 70% To 30% Over The Past Year

A Bloomberg chart drawing on OpenRouter and Exponential View data puts the combined Google, OpenAI and Anthropic share of tokens routed through the platform near 70% in June 2025 and roughly 30% a year later.

Published 2026-07-06Source: OfficeChai
Generated Tokenmaxxing editorial thumbnail for Share Of US Models Being Used On OpenRouter Has Collapsed From 70% To 30% Over The Past Year

Why it matters

OpenRouter meters real production traffic, so its board reads more like a spend report than a benchmark. DeepSeek alone now takes 16.3% of token volume, ahead of Google, Anthropic and OpenAI, while OpenAI has slipped to fourth by company volume.

Tokenmaxxing read

Tencent's Hy3 Preview lists at $0.063 per million input tokens, and that spread is the story: routing defaults move long before the benchmark argument is settled. Amodei's rebuttal — distilled, benchmark-tuned rivals — is about capability, not about what teams default to.

Source takeaway

Claude is the exception inside its own downturn: Opus 4.7 still holds third place and the top closed-source slot, even as six Chinese open-weight models fill out the top ten and Ramp's June report shows DeepSeek leading paid foundational-LLM spend.

Topic links

tokenmaxxingmodel-routingtopicscoreboardsopen-weightspricingtopic
Related projects

Tools that match this angle

#1Direct
Routing

LiteLLM

BerriAI/litellm

An OpenAI-compatible gateway and SDK for calling many model providers with budgets, logging, load balancing, guardrails, and cost tracking.

56.5K10.7KSource-available
gatewaycost-trackingrouting
#2Direct
Observability

Langfuse

langfuse/langfuse

Open-source LLM engineering platform for observability, traces, metrics, evals, prompt management, datasets, and playground workflows.

33.2K3.6KSource-available
tracesevalscosts
#10Direct
Routing

Portkey Gateway

Portkey-AI/gateway

An AI gateway for routing across LLMs with guardrails, provider abstraction, and an OpenAI-compatible API surface.

12.7K1.2KMIT
gatewayguardrailsrouting
Related feed

More source-linked context

Generated Tokenmaxxing editorial thumbnail for Meituan open-sources LongCat-2.0 — the 1.6T model that topped OpenRouter as Owl Alpha
newsW
news

Meituan open-sources LongCat-2.0 — the 1.6T model that topped OpenRouter as Owl Alpha

WinBuzzer: Meituan opened LongCat-2.0, a 1.6-trillion-parameter MoE coding model (~48B active per token, 1M-token context) that surfaced atop OpenRouter as the unbranded alias Owl Alpha — MIT-licensed, with weights not yet posted.

tokenmaxxingmodel-routermodel-routing
Read note
TechNode source artwork
newsT
news

DeepSeek V4 Flash tops OpenRouter weekly ranking with 7.22 trillion tokens · TechNode

DeepSeek V4 Flash led OpenRouter's July 27 to Aug. 2 usage ranking with 7.22 trillion tokens. Chinese models held all four leading slots, and V4 Flash 0731 plus V4 Pro landed inside the top six.

tokenmaxxingmodel-routerpricing
Read note
SaaStrAI source artwork
long-formS
long-form

20VC x SaaStr This Week : Apple Sues OpenAI, the Token-Maxing Era Begins, and the TAM Question Hanging Over AI Coding

On the 20VC/SaaStr podcast, Jason Lemkin, Harry Stebbings and Rory O'Driscoll call the budget era: ClickHouse has grown its AI bill 60-fold since February, while the best engineers keep 10 to 20 agents busy overnight.

tokenmaxxingai-spendmodel-routing
Read note