Claude Haiku 5.5 vs GPT-5.4
Claude Haiku 5.5 leads GPT-5.4 by 5.9 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Haiku 5.5 wins 6 and GPT-5.4 wins 11 of 17 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Haiku 5.5 | GPT-5.4 |
|---|---|---|
| Agentic coding | 67 | 57 |
| Long-horizon autonomy | 62 | 61 |
| Knowledge-work agents | 46 | 40 |
| Computer use | 47 | 42 |
| Security & context learning | 55 | 54 |
Shared benchmarks
| Benchmark | Claude Haiku 5.5 | GPT-5.4 |
|---|---|---|
| APEX-Agentsindex | 56.1% | 52.4% |
| Vending-Bench 2index | $3,602 | $6,144 |
| ARC-AGI | 89.0% | 93.7% |
| ARC-AGI-2 | 61.3% | 74.0% |
| Chess Puzzles | 55.0% | 48.0% |
| CritPt | 22.6% | 23.4% |
| DTBench | 95.0% | 94.4% |
| FrontierMath Tier 4 | 46.3% | 49.0% |
| FrontierMath Tiers 1-3 | 75.1% | 78.6% |
| GPQA Diamond | 89.6% | 93.3% |
| Humanity's Last Exam | 30.1% | 36.2% |
| LMCA | 51.1% | 52.0% |
| MirrorCode | 78.9% | 23.3% |
| OTIS Mock AIME | 98.9% | 97.8% |
| ProofBench | 89.0% | 56.0% |
| SciCode | 55.0% | 56.6% |
| SimpleQA Verified | 23.8% | 45.1% |
Specs and price
| Claude Haiku 5.5 | GPT-5.4 | |
|---|---|---|
| Agent Index | 57.5 (#18) | 51.6 (#33) |
| Agent $/M | $0.050 | $1.27 |
| Input / output $/M | $0.10 / $0.50 | $2.50 / $15.00 |
| Context | 1M | 1.1M |
| Released | Oct 7, 2026 | Mar 5, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 4 |
Don’t pick one. Route.
Shogo can run Claude Haiku 5.5 and GPT-5.4 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo