Claude 3 OpusvsClaude 3.5 Haiku
Claude 3 Opus | Claude 3.5 Haiku | |
|---|---|---|
| Benchmarks | ||
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. | 33% | 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. | 50.4% | 41.6% |
| BenchmarksPublished by one model only | ||
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% |
| Overview | ||
| Company | Anthropic | Anthropic |
| Release date | Mar 4 2024 | Oct 22 2024 |
| Access | Proprietary | Proprietary |
Other comparisons
Claude 3 OpusvsGPT-6 AstraClaude 3.5 HaikuvsGPT-6 AstraClaude 3 OpusvsGemini 3.8 FlashClaude 3.5 HaikuvsGemini 3.8 FlashClaude 3 OpusvsMuse Spark 1.3Claude 3.5 HaikuvsMuse Spark 1.3Claude 3 OpusvsGrok 4.6Claude 3.5 HaikuvsGrok 4.6Claude 3 OpusvsDeepSeek-V4.1-FlashClaude 3.5 HaikuvsDeepSeek-V4.1-FlashClaude 3 OpusvsMistral Medium 3.5Claude 3.5 HaikuvsMistral Medium 3.5
Frequently asked questions
Claude 3 Opus and Claude 3.5 Haiku are evenly matched across the 2 benchmarks they both report (SWE-Bench Verified, GPQA Diamond). Claude 3 Opus shipped 232 days before Claude 3.5 Haiku, 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 SWE-Bench Verified, Claude 3.5 Haiku leads at 40.6% vs Claude 3 Opus at 33%. On GPQA Diamond, Claude 3 Opus leads at 50.4% vs Claude 3.5 Haiku at 41.6%.