Claude Sonnet 5 vs Gemini 3.8 Flash
| Spec | Claude Sonnet 5 | Gemini 3.8 Flash |
|---|---|---|
| Respan Index | 46.7 (#39) | 54.6 (#22) |
| Blended price per 1M | $4 | $0.75 |
| Context window | 1M | 1.0M |
| Input / output per 1M | $2 / $10 | $0.375 / $1.88 |
| Max output | 128K | 66K |
| Input | text, vision | text, vision, audio |
| Released | 2026-06-30 | 2026-09-02 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Claude Sonnet 5 | Gemini 3.8 Flash |
|---|---|---|
| Coding | 26.9 Claude Sonnet 5 | 36.7 Gemini 3.8 Flash |
| Agents | 62.8 Claude Sonnet 5 | 66.2 Gemini 3.8 Flash |
| Reasoning | 51.0 Claude Sonnet 5 | 53.7 Gemini 3.8 Flash |
| Knowledge & language | 34.0 Claude Sonnet 5 | 70.9 Gemini 3.8 Flash |
| Human preference | 54.4 Claude Sonnet 5 | 55.8 Gemini 3.8 Flash |
Benchmarks both report
Gemini 3.8 Flash leads on 17 of the 25 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Claude Sonnet 5 | Gemini 3.8 Flash |
|---|---|---|
| Chess Puzzles (Epoch AI) | 45% | 58% |
| LMArena Agent (LMArena) | 0.0294 | 0.0265 |
| LMArena Elo | 1462 | 1496 |
| LMArena WebDev (LMArena) | 1540 | 1583 |
| LMArena Vision (LMArena) | 1262 | 1291 |
| Terminal-Bench 4.0 (Vals AI) | 10.6% | 19.7% |
| APEX-Agents (Mercor) | 54.5% | 64.3% |
| DeepSWE v1.1 (Datacurve) | 53.8% | 73.8% |
| FrontierMath Tier 4 v2 (Epoch AI) | 29.3% | 22% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 65.6% | 68.4% |
| GPQA Diamond | 90.5% | 95.4% |
| LiveBench Agentic Coding (LiveBench) | 59.4% | 54.2% |
| LiveBench Coding (LiveBench) | 80.7% | 72.5% |
| LiveBench Data Analysis (LiveBench) | 71.7% | 54% |
| LiveBench Instruction Following (LiveBench) | 63.9% | 81.4% |
| LiveBench Language (LiveBench) | 75% | 87.8% |
| LiveBench Mathematics (LiveBench) | 92.9% | 91.6% |
| LiveBench Reasoning (LiveBench) | 88.7% | 89.3% |
| Mystery Game Puzzles (Epoch AI) | 35% | 47% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 94.7% | 98.9% |
| SAGE (Vals AI) | 48.9% | 35.1% |
| SimpleQA Verified | 33.7% | 69.7% |
| Vending-Bench 2 (Andon Labs) | 6377.7 | 5093.79 |
| WeirdML (Håvard Tveit Ihle) | 68.8% | 84.8% |
| SimpleBench (SimpleBench) | 60.6% | 82.4% |
FAQ
Is Claude Sonnet 5 better than Gemini 3.8 Flash?
Gemini 3.8 Flash scores 54.6 on the Respan Index (#22), ahead of Claude Sonnet 5 at 46.7 (#39). Gemini 3.8 Flash leads on 17 of the 25 benchmarks both report.
Which is cheaper, Claude Sonnet 5 or Gemini 3.8 Flash?
Claude Sonnet 5 lists at $2 input and $10 output per 1M tokens; Gemini 3.8 Flash at $0.375 and $1.88. Gemini 3.8 Flash is 5.3x cheaper per token at list price.
Which has the larger context window?
Claude Sonnet 5 accepts 1M tokens and Gemini 3.8 Flash 1.0M tokens.