Claude Sonnet 5 vs Gemini 3.8 Flash
Gemini 3.8 Flash leads Claude Sonnet 5 by 3.8 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Sonnet 5 wins 5 and Gemini 3.8 Flash wins 15 of 20 shared benchmarks. Gemini 3.8 Flash is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Sonnet 5 | Gemini 3.8 Flash |
|---|---|---|
| Agentic coding | 57 | 63 |
| Long-horizon autonomy | 65 | 64 |
| Knowledge-work agents | 43 | 47 |
| Computer use | 40 | 44 |
| Security & context learning | 51 | 53 |
Shared benchmarks
| Benchmark | Claude Sonnet 5 | Gemini 3.8 Flash |
|---|---|---|
| APEX-Agentsindex | 54.5% | 64.3% |
| DeepSWEindex | 53.8% | 73.8% |
| Vending-Bench 2index | $6,378 | $5,094 |
| CursorBenchindex | 34.1% | 39.6% |
| Chess Puzzles | 45.0% | 58.0% |
| CritPt | 16.9% | 18.3% |
| DTBench | 92.5% | 95.7% |
| FrontierCode | 42.7% | 41.2% |
| FrontierMath Tier 4 | 29.3% | 22.0% |
| FrontierMath Tiers 1-3 | 65.6% | 68.4% |
| GeoBench | 74.0% | 86.0% |
| GPQA Diamond | 90.5% | 95.4% |
| LMCA | 50.0% | 52.9% |
| MirrorCode | 63.3% | 38.9% |
| OTIS Mock AIME | 94.7% | 98.9% |
| ProofBench | 77.0% | 48.0% |
| SciCode | 54.3% | 56.6% |
| SimpleBench | 60.6% | 82.4% |
| SimpleQA Verified | 33.7% | 69.7% |
| WebDev Arena | 1537 | 1568 |
Specs and price
| Claude Sonnet 5 | Gemini 3.8 Flash | |
|---|---|---|
| Agent Index | 52.4 (#32) | 56.2 (#21) |
| Agent $/M | $0.92 | $0.35 |
| Input / output $/M | $2.00 / $10.00 | $0.75 / $3.75 |
| Context | 1M | 1M |
| Released | Jun 30, 2026 | Sep 2, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 6 |
Don’t pick one. Route.
Shogo can run Claude Sonnet 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