GPT-5.6 Sol vs GPT-5.6 Terra
GPT-5.6 Sol leads GPT-5.6 Terra by 3.0 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. GPT-5.6 Sol wins 20 and GPT-5.6 Terra wins 2 of 22 shared benchmarks. GPT-5.6 Terra is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | GPT-5.6 Sol | GPT-5.6 Terra |
|---|---|---|
| Agentic coding | 71 | 66 |
| Long-horizon autonomy | 72 | 69 |
| Knowledge-work agents | 46 | 47 |
| Computer use | 53 | 48 |
| Security & context learning | 58 | 56 |
Shared benchmarks
| Benchmark | GPT-5.6 Sol | GPT-5.6 Terra |
|---|---|---|
| APEX-Agentsindex | 51.4% | 58.2% |
| DeepSWEindex | 72.7% | 69.6% |
| Vending-Bench 2index | $9,619 | $7,343 |
| CursorBenchindex | 41.7% | 41.3% |
| ARC-AGI | 97.5% | 96.5% |
| ARC-AGI-2 | 92.5% | 83.9% |
| BALROG | 60.0% | 53.2% |
| Chess Puzzles | 68.0% | 47.0% |
| CritPt | 32.3% | 30.0% |
| DTBench | 96.0% | 93.3% |
| FrontierCode | 47.5% | 41.3% |
| FrontierMath Tier 4 | 82.9% | 70.7% |
| FrontierMath Tiers 1-3 | 89.1% | 86.0% |
| GPQA Diamond | 93.5% | 93.3% |
| LMCA | 59.2% | 54.9% |
| MirrorCode | 20.0% | 41.1% |
| OTIS Mock AIME | 100.0% | 99.7% |
| ProofBench | 83.0% | 74.0% |
| SciCode | 57.8% | 55.0% |
| SimpleBench | 71.7% | 48.9% |
| SimpleQA Verified | 69.7% | 43.2% |
| WebDev Arena | 1617 | 1521 |
Specs and price
| GPT-5.6 Sol | GPT-5.6 Terra | |
|---|---|---|
| Agent Index | 61.8 (#14) | 58.8 (#16) |
| Agent $/M | $1.85 | $1.02 |
| Input / output $/M | $4.00 / $20.00 | $2.00 / $12.00 |
| Context | 1.1M | 1.1M |
| Released | Aug 7, 2026 | Jul 9, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 4 |
Don’t pick one. Route.
Shogo can run GPT-5.6 Sol 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