cd /news/large-language-models/from-gpt-5-can-t-do-basic-math-to-to… · home › topics › large-language-models › article
[ARTICLE · art-83508] src=promptcube3.com ↗ pub= topic=large-language-models verified=true sentiment=· neutral

From 'GPT-5 Can't Do Basic Math' to Today: What a Year Tells Us

A year after GPT-5's release, its mathematical and reasoning capabilities have improved dramatically, with users reporting it can now walk through proofs, though some note real-world reliability issues persist on simple logic tasks.

read1 min views1 publishedAug 2, 2026
From 'GPT-5 Can't Do Basic Math' to Today: What a Year Tells Us
Image: Promptcube3 (auto-discovered)

A year ago

Next Figure AI's F.03 Ladder Climb: Autonomy or OSHA Violation? →

All Replies （3） #

C

A year ago it couldn't add fractions reliably. Now it walks through proofs for me. Night and day.

0

Z

Benchmarks improved, sure, but real-world reliability still stumbles on simple logic. Progress ≠ parity.

0

R

How are you catching those logic stumbles in your own testing? Curious what eval setup you use.

0

source & further reading
promptcube3.com — original article Figure AI's F.03 Ladder Climb: Autonomy or OSHA Violation? OpenAI's Connes Rigidity Disproof Looks Wrong on Inspection Title: CostPerPrompt: Live AI API Pricing & Real Cost Calculator
── more in #large-language-models 4 stories · sorted by recency
── more on @gpt-5 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/from-gpt-5-can-t-do-…] indexed:0 read:1min 2026-08-02 · —