← sources

I think Claude is going to lose the AI battle to ChatGPT

source: reddit (reddit:AI_Agents)published 2026-09-10 17:45ingested 2026-09-10 22:38

retained text

I have been looking at the latest Claude vs ChatGPT numbers, and the picture is a lot more interesting than X is better than Y. ( frr no body is better i believe opensource all the way )

but yeah GPT-6 Astra is ahead of Claude Fable 5.1 on several major benchmarks:

• Terminal-Bench 4.0: **57.9% vs 55.8%**  
• DeepSWE: **74.1% vs 67.4%**  
• FrontierMath Tier 4: **97.6% vs 87.8%**  
• GPQA Diamond: **96.0% vs 93.7%**  
• ARC-AGI-3: **99.9% vs zero**  
• OSWorld 2.0: **72.6% vs 70.2%**

But the data isn't a clean sweep. , Claude Fable 5.1 scores **65.0% vs Astra’s 57.2%**.

And on independent evaluations, Claude still performs extremely competitively.

The user numbers tell another story:

ChatGPT had roughly **1.11B monthly users** in May, compared with **245M for Claude**. But Claude's growth was much faster, rising from **60.2M to 245M** in roughly five months.

Enterprise spending also tells a different story, with some estimates putting Claude ahead of OpenAI in enterprise LLM spend.

So I don't think the data supports the idea that Claude is finished.

What it supports is something more interesting:

**Astra appears to have moved the benchmark race significantly toward OpenAI, while Anthropic is still gaining users and enterprise share.**

The next question isn't who has the highest benchmark score.

It's whether Astra's capability advantage translates into **sustained user growth, developer adoption and enterprise revenue.**

That's the part we don't have enough data to call yet.

document fetched 2026-09-11 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1wcq94z/i_think_claude_is_going_to_lose_the_ai_battle_to