GPT-5.4 (2026-03-05) vs GPT-5.5
EBR Bench?
GPT-5.4 (2026-03-05)at xhigh25.4
GPT-5.5at xhigh34.3
Toolathlon-Verified?
GPT-5.4 (2026-03-05)—
GPT-5.5at xhigh73.5
MCP-Atlas Public?
GPT-5.4 (2026-03-05)—
GPT-5.5at xhigh79.4
MCP-Atlas · full 1000?
GPT-5.4 (2026-03-05)—
GPT-5.5at xhigh75.3
| Benchmark | GPT-5.4 (2026-03-05) | GPT-5.5 |
|---|---|---|
| AutomationBench-AA | — | 40.1 |
| EBR Bench | 25.4 · xhigh | 34.3 · xhigh |
| Furniture Assembly | 37.5 · xhigh | 44.2 · xhigh |
| GDPval-AA (work tasks) 2.1 | — | 1,231.9 |
| Toolathlon-Verified | — | 73.5 · xhigh |
| MCP-Atlas Public | — | 79.4 · xhigh |
| MCP-Atlas · full 1000 | — | 75.3 · xhigh |
| τ³-bench · banking · 1.0.1 | — | 44.6 · xhigh |
| τ³-Banking · AA | — | 29.9 |
| AA Analyst Agent | — | 50 · xhigh |
| ITBench SRE | — | 45.8 · xhigh |
| APEX-Agents | — | 37.7 · xhigh |
| Terminal-Bench 4.0 | — | 5.1 |
| Terminal-Bench Hard (version unstated) | — | 57.6 |
| AA-Briefcase 1.1 · Elo | — | 1,000 |
| AA-LCR (long context) 1.1 | — | 83 |
| DeepSWE 1.1 | — | 54 |
| EnterpriseOps-Gym-AA | — | 44.1 |
Specs
| Spec | GPT-5.4 (2026-03-05) | GPT-5.5 |
|---|---|---|
| Input, per 1M | !$2.50 | $5 |
| Output, per 1M | !$15 | $30 |
| Cost per task? | — | $0.77 |
| Speed, tok/s | — | 82 |
| Context | 1.05M | 1.05M |
| Weights | Not verified | Closed |
| Released | Mar 2026 | Apr 2026 |
| See images | ✓ | ✓ |
| Read PDFs | ? | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | ? | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
GPT-5.4 (2026-03-05) → GPT-5.5
Prompts with >272K input tokens›
Above 272K input tokens, input costs 2× and output 1.5× for the full session in Standard, Batch and Flex.
GPT-5.5 → GPT-5.4 (2026-03-05)
Nothing known.