GPT-5.5 vs GPT-5.6 Terra
GPT-5.6 Terra leads GPT-5.5 by 2.1 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. GPT-5.5 wins 11 and GPT-5.6 Terra wins 9 of 20 shared benchmarks. GPT-5.6 Terra is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | GPT-5.5 | GPT-5.6 Terra |
|---|---|---|
| Agentic coding | 63 | 66 |
| Long-horizon autonomy | 67 | 69 |
| Knowledge-work agents | 46 | 47 |
| Computer use | 43 | 48 |
| Security & context learning | 55 | 56 |
Shared benchmarks
| Benchmark | GPT-5.5 | GPT-5.6 Terra |
|---|---|---|
| APEX-Agentsindex | 55.1% | 58.2% |
| DeepSWEindex | 67.0% | 69.6% |
| Vending-Bench 2index | $7,524 | $7,343 |
| ARC-AGI | 95.0% | 96.5% |
| ARC-AGI-2 | 85.0% | 83.9% |
| Chess Puzzles | 63.0% | 47.0% |
| CritPt | 27.1% | 30.0% |
| DTBench | 96.0% | 93.3% |
| FrontierCode | 43.0% | 41.3% |
| FrontierMath Tier 4 | 72.5% | 70.7% |
| FrontierMath Tiers 1-3 | 85.3% | 86.0% |
| GPQA Diamond | 94.0% | 93.3% |
| LMCA | 54.3% | 54.9% |
| MirrorCode | 23.3% | 41.1% |
| OTIS Mock AIME | 100.0% | 99.7% |
| ProofBench | 50.0% | 74.0% |
| SciCode | 56.1% | 55.0% |
| SimpleBench | 69.0% | 48.9% |
| SimpleQA Verified | 63.0% | 43.2% |
| WebDev Arena | 1510 | 1521 |
Specs and price
| GPT-5.5 | GPT-5.6 Terra | |
|---|---|---|
| Agent Index | 56.7 (#19) | 58.8 (#16) |
| Agent $/M | $2.55 | $1.02 |
| Input / output $/M | $5.00 / $30.00 | $2.00 / $12.00 |
| Context | 1.1M | 1.1M |
| Released | Apr 23, 2026 | Jul 9, 2026 |
| Weights | Closed | Closed |
| API hosts | 4 | 4 |
Don’t pick one. Route.
Shogo can run GPT-5.5 and GPT-5.6 Terra side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo