GPT-5.5
AI models review
As of Aug 12, 2026, GPT-5.5 is rated 1481 Elo on LM Arena (55,130 Arena votes).
Benchmark and price facts
- Vendor
- OpenAI
- License
- proprietary
- LM Arena text
- 1481 Elo (#17 of 20)
- Price (input)
- $5 / 1M tokens
- Price (output)
- $30 / 1M tokens
- Context window
- 1,050,000 tokens
- AA Intelligence Index
- 54.7 / 100
- AA Coding Index
- 71.6 / 100
- Human votes
- 55,130
Elo and votes from LM Arena; intelligence, coding and speed indices from Artificial Analysis; price and context window from OpenRouter. Real fetched values, attributed, never blended.
Is GPT-5.5 worth it in 2026?
GPT-5.5 sits at rank 17 of 20 in LM Arena text Elo with a score of 1481 from about 55,000 votes, placing it near the bottom of this competitive group for general chat quality. Its Coding Index of 71.6 outpaces its general Intelligence Index of 54.7, suggesting coding is its relative strength. The 1,050,000-token context window is a standout practical feature for long-document work. However, output pricing at 30 USD per 1M tokens is steep, making high-volume use costly.
Coding tasks and long-document processing where the large context window is needed and cost is not the primary concern.
At rank 17 of 20 for text Elo and 30 USD per 1M output tokens, users pay a premium price for below-average conversational quality within this benchmark group.
AI-generated editorial synthesis of the source scores and sampled review themes, not a hands-on test. Reviewer experiences and allegations are not independently verified facts.
Rated near GPT-5.5
Closest scores in ai models, on the same source.