Model comparison
GPT-5.5 Pro vs Sonar
GPT-5.5 Pro is the stronger model overall, scoring 64.3 to 38.5 on the Noometry Index. Sonar costs 68× less per token, which makes it the better buy when GPT-5.5 Pro's lead doesn't matter for your workload.
Last verified . 0 shared benchmarks.
PerplexitySummary
- The widest gap is in reasoning, where GPT-5.5 Pro leads 73.3 to 21.1.
- Sonar is cheaper at $1 / $1 per million input/output tokens, against $30 / $180 for GPT-5.5 Pro.
- GPT-5.5 Pro accepts more context: 1.05M tokens versus 128K.
Side by side
| GPT-5.5 Pro | Sonar | |
|---|---|---|
| Provider | OpenAI | Perplexity |
| Noometry Index | 64.3 | 38.5 |
| Released | 2026-04-23 | 2024-01-01 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 128K |
| Max output | 128K | 4K |
| Input $ / M tokens | $30 | $1 |
| Output $ / M tokens | $180 | $1 |
| Results tracked | 14 | 7 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
GPT-5.5 Pro: —, Sonar: 35.7 (#221)
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| LiveBench Coding | — | 35.1% |
Reasoning GPT-5.5 Pro leads
GPT-5.5 Pro: 73.3 (#10), Sonar: 21.1 (#227)
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| ARC-AGI-2 | 84.6% | — |
| SimpleBench | 76.9% | — |
| ARC-AGI-1 | 96.5% | — |
| CritPt | 30.6% | — |
| Chess Puzzles | 64% | — |
| LiveBench Reasoning | — | 46.3% |
| DTBench | 96% | — |
| LiveBench Data Analysis | — | 37.9% |
| LMCA | 53.9% | — |
| Epoch Capabilities Index | 162.07 | — |
| LiveBench | — | 46.9% |
Math GPT-5.5 Pro leads
GPT-5.5 Pro: 84.0 (#10), Sonar: 33.7 (#200)
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| FrontierMath (Tiers 1-3) | 87.7% | — |
| FrontierMath Tier 4 | 78% | — |
| OTIS Mock AIME 2024-2025 | 100% | — |
| LiveBench Math | — | 41.6% |
| FrontierMath (Feb 2025 set) | 52.4% | — |
| FrontierMath Tier 4 (v1) | 39.6% | — |
Knowledge Not comparable
GPT-5.5 Pro: 64.1 (#19), Sonar: —
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| GPQA Diamond | 93.9% | — |
Instruction Following Not comparable
GPT-5.5 Pro: —, Sonar: 71.4 (#150)
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| LiveBench Instruction Following | — | 76.2% |
Writing & Preference Not comparable
GPT-5.5 Pro: —, Sonar: 52.6 (#167)
| Benchmark | GPT-5.5 Pro | Sonar |
|---|---|---|
| LiveBench Language | — | 44.1% |
Frequently asked questions
Is GPT-5.5 Pro better than Sonar?
GPT-5.5 Pro is the stronger model overall, scoring 64.3 to 38.5 on the Noometry Index. Sonar costs 68× less per token, which makes it the better buy when GPT-5.5 Pro's lead doesn't matter for your workload.
Which is cheaper, GPT-5.5 Pro or Sonar?
Sonar is cheaper. It lists at $1 per million input tokens and $1 per million output tokens; GPT-5.5 Pro lists at $30 and $180.
Which has the bigger context window?
GPT-5.5 Pro does, with 1.05M tokens against 128K.
How many benchmarks do GPT-5.5 Pro and Sonar share?
0 benchmarks have published results for both models. GPT-5.5 Pro has 14 scored results on Noometry and Sonar has 7.