Settle it race by race.
Overall scores hide trade-offs. Line models up to see who wins each race, by how much, and what you pay for the difference.
Verdict
Opus 5.5 leads the overall intelligence race by 16.8 pts
Dominant leadover GPT-6.1 Sol. Race wins below decide most real choices.
Coding, Hard Reasoning, Math, Science, Writing & Language, Multimodal, Long Context
Speed & Cost, Trust & Factuality
Capability shape
Scores normalised to 0 to 100 across language races. Hover a name to isolate it.
Where Opus 5.5 wins biggest
- Writing & Language47.9 ptsDominant lead
- Multimodal Understanding38.3 ptsDominant lead
- Science30.5 ptsDominant lead
- Overall Intelligence16.8 ptsDominant lead
- Coding14.3 ptsDominant lead
Where GPT-6.1 Sol wins biggest
- Time to First Token6.08sDominant lead
- Speed & Cost7.3 ptsDominant lead
- Graduate Science QA4.8 ptsDominant lead
- Trust & Factuality4.3 ptsDominant lead
- Intelligence per Dollar4.1 ptsDominant lead
Race by race
Dots sit at true distances. The white tick is the best score anyone has right now.
- Opus 5.591.1#1GPT-6.1 Sol74.3#8
- Opus 5.598.3#1GPT-6.1 Sol84.0#6
- Opus 5.597.9#1GPT-6.1 Sol93.1#4
- Opus 5.592.4#2GPT-6.1 Sol81.9#7
- Opus 5.585.5#5GPT-6.1 Sol55.0#27
- Opus 5.597.3#2GPT-6.1 Sol49.4#35
- Opus 5.597.9#1GPT-6.1 Sol59.6#19
- No scoresOpus 5.5n/aGPT-6.1 Soln/a
- Opus 5.595.7#4GPT-6.1 Sol90.6#6
- Opus 5.536.3#22GPT-6.1 Sol43.6#18
- Opus 5.591.5#5GPT-6.1 Sol95.7#2
Overall Intelligence over time
Weekly points since January 2025 where the source has history; as-measured races show one point per model. Hover for exact scores.
Specs and pricing
Best value in each row is highlighted.
| Opus 5.5 | GPT-6.1 Sol | |
|---|---|---|
| Released | Sep 22, 2026 | Sep 29, 2026 |
| Weights | Closed | Closed |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Speed | 96.4 tok/s | 55.1 tok/s |
| First token | 7.56s | 1.48s |
| Input per 1M tokens | $4 | $2 |
| Output per 1M tokens | $20 | $10 |
| Inputs | text, image | image, text |
| Outputs | text | text |