Claude Haiku 5.5 vs Gemini 3.8 Flash
Claude Haiku 5.5 leads Gemini 3.8 Flash by 1.3 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Haiku 5.5 wins 9 and Gemini 3.8 Flash wins 12 of 22 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Haiku 5.5 | Gemini 3.8 Flash |
|---|---|---|
| Agentic coding | 67 | 63 |
| Long-horizon autonomy | 62 | 64 |
| Knowledge-work agents | 46 | 47 |
| Computer use | 47 | 44 |
| Security & context learning | 55 | 53 |
Shared benchmarks
| Benchmark | Claude Haiku 5.5 | Gemini 3.8 Flash |
|---|---|---|
| APEX-Agentsindex | 56.1% | 64.3% |
| Vending-Bench 2index | $3,602 | $5,094 |
| FrontierSWEindex | 43.8% | 19.6% |
| CursorBenchindex | 48.4% | 39.6% |
| ARC-AGI | 89.0% | 98.5% |
| ARC-AGI-2 | 61.3% | 89.2% |
| Chess Puzzles | 55.0% | 58.0% |
| CritPt | 22.6% | 18.3% |
| DTBench | 95.0% | 95.7% |
| FrontierCode | 46.4% | 41.2% |
| FrontierMath Tier 4 | 46.3% | 22.0% |
| FrontierMath Tiers 1-3 | 75.1% | 68.4% |
| GPQA Diamond | 89.6% | 95.4% |
| Humanity's Last Exam | 30.1% | 44.5% |
| LMCA | 51.1% | 52.9% |
| MirrorCode | 78.9% | 38.9% |
| OTIS Mock AIME | 98.9% | 98.9% |
| ProofBench | 89.0% | 48.0% |
| SciCode | 55.0% | 56.6% |
| SimpleBench | 59.2% | 82.4% |
| SimpleQA Verified | 23.8% | 69.7% |
| WeirdML v3 | 11.7% | 7.4% |
Specs and price
| Claude Haiku 5.5 | Gemini 3.8 Flash | |
|---|---|---|
| Agent Index | 57.5 (#18) | 56.2 (#21) |
| Agent $/M | $0.050 | $0.35 |
| Input / output $/M | $0.10 / $0.50 | $0.75 / $3.75 |
| Context | 1M | 1M |
| Released | Oct 7, 2026 | Sep 2, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 6 |
Don’t pick one. Route.
Shogo can run Claude Haiku 5.5 and Gemini 3.8 Flash side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo