Compare AI models
| Benchmarks | ||
BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | 50% | 39% |
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 40.6% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 41.6% | — |
| Overview | ||
| Company | Anthropic | OpenAI |
| Release date | Oct 22 2024 | Sep 15 2025 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Claude 3.5 HaikuvsGemini 4 ArgonGPT-5-CodexvsGemini 4 ArgonClaude 3.5 HaikuvsMuse WonderGPT-5-CodexvsMuse WonderClaude 3.5 HaikuvsGrok 4.7GPT-5-CodexvsGrok 4.7Claude 3.5 HaikuvsDeepSeek-V4.1-FlashGPT-5-CodexvsDeepSeek-V4.1-FlashClaude 3.5 HaikuvsMistral Large 4GPT-5-CodexvsMistral Large 4Claude 3.5 HaikuvsKimi K3GPT-5-CodexvsKimi K3Frequently asked questions
Claude 3.5 Haiku leads GPT-5-Codex on 1 of the 1 benchmark they both report (BullshitBench v2). Claude 3.5 Haiku shipped 328 days before GPT-5-Codex, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On BullshitBench v2, Claude 3.5 Haiku leads at 50% vs GPT-5-Codex at 39%.