Claude Haiku 5.5 vs GPT-6 Sol
GPT-6 Sol leads Claude Haiku 5.5 by 6.1 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Haiku 5.5 wins 3 and GPT-6 Sol wins 16 of 19 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Haiku 5.5 | GPT-6 Sol |
|---|---|---|
| Agentic coding | 67 | 72 |
| Long-horizon autonomy | 62 | 76 |
| Knowledge-work agents | 46 | 48 |
| Computer use | 47 | 54 |
| Security & context learning | 55 | 59 |
Shared benchmarks
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol |
|---|---|---|
| APEX-Agentsindex | 56.1% | 54.3% |
| Vending-Bench 2index | $3,602 | $14,428 |
| ARC-AGI | 89.0% | 95.5% |
| ARC-AGI-2 | 61.3% | 89.6% |
| Chess Puzzles | 55.0% | 56.0% |
| CritPt | 22.6% | 30.9% |
| DTBench | 95.0% | 97.3% |
| FrontierCode | 46.4% | 49.3% |
| FrontierMath Tier 4 | 46.3% | 90.0% |
| FrontierMath Tiers 1-3 | 75.1% | 89.8% |
| GPQA Diamond | 89.6% | 94.3% |
| LMCA | 51.1% | 59.1% |
| MirrorCode | 78.9% | 48.9% |
| OTIS Mock AIME | 98.9% | 100.0% |
| ProofBench | 89.0% | 83.0% |
| SciCode | 55.0% | 57.6% |
| SimpleBench | 59.2% | 73.1% |
| SimpleQA Verified | 23.8% | 60.7% |
| WeirdML v3 | 11.7% | 19.7% |
Specs and price
| Claude Haiku 5.5 | GPT-6 Sol | |
|---|---|---|
| Agent Index | 57.5 (#18) | 63.6 (#13) |
| Agent $/M | $0.050 | $0.92 |
| Input / output $/M | $0.10 / $0.50 | $2.00 / $10.00 |
| Context | 1M | 1.1M |
| Released | Oct 7, 2026 | Sep 22, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 5 |
Don’t pick one. Route.
Shogo can run Claude Haiku 5.5 and GPT-6 Sol side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo