Claude Opus 5 vs Gemini 3.8 Flash
Claude Opus 5 leads Gemini 3.8 Flash by 9.2 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Opus 5 wins 15 and Gemini 3.8 Flash wins 7 of 23 shared benchmarks. Gemini 3.8 Flash is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Opus 5 | Gemini 3.8 Flash |
|---|---|---|
| Agentic coding | 74 | 63 |
| Long-horizon autonomy | 74 | 64 |
| Knowledge-work agents | 52 | 47 |
| Computer use | 56 | 44 |
| Security & context learning | 59 | 53 |
Shared benchmarks
| Benchmark | Claude Opus 5 | Gemini 3.8 Flash |
|---|---|---|
| APEX-Agentsindex | 65.8% | 64.3% |
| DeepSWEindex | 73.6% | 73.8% |
| Vending-Bench 2index | $11,182 | $5,094 |
| FrontierSWEindex | 52.0% | 19.6% |
| CursorBenchindex | 46.6% | 39.6% |
| ARC-AGI | 97.5% | 98.5% |
| ARC-AGI-2 | 90.4% | 89.2% |
| Chess Puzzles | 43.0% | 58.0% |
| CritPt | 29.1% | 18.3% |
| DTBench | 97.9% | 95.7% |
| FrontierCode | 53.4% | 41.2% |
| FrontierMath Tier 4 | 73.2% | 22.0% |
| FrontierMath Tiers 1-3 | 85.6% | 68.4% |
| GPQA Diamond | 93.9% | 95.4% |
| LMCA | 64.5% | 52.9% |
| MirrorCode | 87.8% | 38.9% |
| OTIS Mock AIME | 98.9% | 98.9% |
| ProofBench | 99.0% | 48.0% |
| SciCode | 56.4% | 56.6% |
| SimpleBench | 80.6% | 82.4% |
| SimpleQA Verified | 59.9% | 69.7% |
| WebDev Arena | 1687 | 1568 |
| WeirdML v3 | 19.0% | 7.4% |
Specs and price
| Claude Opus 5 | Gemini 3.8 Flash | |
|---|---|---|
| Agent Index | 65.4 (#11) | 56.2 (#21) |
| Agent $/M | $2.31 | $0.35 |
| Input / output $/M | $5.00 / $25.00 | $0.75 / $3.75 |
| Context | 1M | 1M |
| Released | Jul 24, 2026 | Sep 2, 2026 |
| Weights | Closed | Closed |
| API hosts | 6 | 6 |
Don’t pick one. Route.
Shogo can run Claude Opus 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