Claude Sonnet 5 vs Gemini 4 Carbon
Gemini 4 Carbon leads Claude Sonnet 5 by 21.2 points on the Shogo Agent Index. Claude Sonnet 5 wins 0 and Gemini 4 Carbon wins 12 of 12 shared benchmarks. Gemini 4 Carbon is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Sonnet 5 | Gemini 4 Carbon |
|---|---|---|
| Agentic coding | 57 | 82 |
| Long-horizon autonomy | 65 | 80 |
| Knowledge-work agents | 43 | 63 |
| Computer use | 40 | 68 |
| Security & context learning | 51 | 63 |
Shared benchmarks
| Benchmark | Claude Sonnet 5 | Gemini 4 Carbon |
|---|---|---|
| APEX-Agentsindex | 54.5% | 84.2% |
| Vending-Bench 2index | $6,378 | $16,484 |
| CursorBenchindex | 34.1% | 60.4% |
| Chess Puzzles | 45.0% | 78.0% |
| CritPt | 16.9% | 36.0% |
| FrontierMath Tier 4 | 29.3% | 95.1% |
| FrontierMath Tiers 1-3 | 65.6% | 90.5% |
| GPQA Diamond | 90.5% | 96.5% |
| OTIS Mock AIME | 94.7% | 100.0% |
| SciCode | 54.3% | 64.2% |
| SimpleBench | 60.6% | 92.6% |
| SimpleQA Verified | 33.7% | 77.7% |
Specs and price
| Claude Sonnet 5 | Gemini 4 Carbon | |
|---|---|---|
| Agent Index | 52.4 (#32) | 73.6 (#2) |
| Agent $/M | $0.92 | $0.84 |
| Input / output $/M | $2.00 / $10.00 | $2.00 / $10.00 |
| Context | 1M | 1M |
| Released | Jun 30, 2026 | Nov 10, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 5 |
Don’t pick one. Route.
Shogo can run Claude Sonnet 5 and Gemini 4 Carbon side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo