Claude Haiku 5.5 vs Grok 4.6
Claude Haiku 5.5 leads Grok 4.6 by 1.3 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Haiku 5.5 wins 10 and Grok 4.6 wins 10 of 20 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Haiku 5.5 | Grok 4.6 |
|---|---|---|
| Agentic coding | 67 | 61 |
| Long-horizon autonomy | 62 | 68 |
| Knowledge-work agents | 46 | 47 |
| Computer use | 47 | 42 |
| Security & context learning | 55 | 52 |
Shared benchmarks
| Benchmark | Claude Haiku 5.5 | Grok 4.6 |
|---|---|---|
| APEX-Agentsindex | 56.1% | 65.3% |
| Vending-Bench 2index | $3,602 | $9,047 |
| FrontierSWEindex | 43.8% | 25.3% |
| CursorBenchindex | 48.4% | 41.4% |
| ARC-AGI | 89.0% | 87.5% |
| ARC-AGI-2 | 61.3% | 67.1% |
| Chess Puzzles | 55.0% | 33.0% |
| CritPt | 22.6% | 19.7% |
| DTBench | 95.0% | 97.3% |
| FrontierCode | 46.4% | 48.0% |
| FrontierMath Tier 4 | 46.3% | 31.7% |
| FrontierMath Tiers 1-3 | 75.1% | 66.0% |
| GPQA Diamond | 89.6% | 94.0% |
| LMCA | 51.1% | 48.5% |
| MirrorCode | 78.9% | 41.1% |
| OTIS Mock AIME | 98.9% | 99.2% |
| ProofBench | 89.0% | 51.0% |
| SciCode | 55.0% | 56.5% |
| SimpleBench | 59.2% | 75.9% |
| SimpleQA Verified | 23.8% | 49.3% |
Specs and price
| Claude Haiku 5.5 | Grok 4.6 | |
|---|---|---|
| Agent Index | 57.5 (#18) | 56.2 (#22) |
| Agent $/M | $0.050 | $0.98 |
| Input / output $/M | $0.10 / $0.50 | $2.00 / $6.00 |
| Context | 1M | 500K |
| Released | Oct 7, 2026 | Aug 12, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 4 |
Don’t pick one. Route.
Shogo can run Claude Haiku 5.5 and Grok 4.6 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo