Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | Nova 2.0 Pro Preview (medium) |
|---|---|---|
| Input ($/M tokens) | $2 | $1.25 |
| Output ($/M tokens) | $10 | $10 |
Verdict. Claude Sonnet 5 (Adaptive Reasoning, Max Effort) wins the overall benchmark matchup 2–0 across 2 overlapping categories, but raw benchmark score is only one input to the decision.
Pricing. Both models sit in the mid-tier bracket for output-token pricing. At 1.0× the per-million-token cost, Nova 2.0 Pro Preview (medium) is meaningfully cheaper if your traffic is output-heavy (long completions, document generation, agent loops). Nova 2.0 Pro Preview (medium) makes more sense when output volume is low and absolute reasoning quality justifies the premium.
Strengths. Claude Sonnet 5 (Adaptive Reasoning, Max Effort) is strongest on Coding Index (71.5), Intelligence Index (55.3). Nova 2.0 Pro Preview (medium) leads on Coding Index (34.0), Intelligence Index (22.1).
Speed. On throughput, Nova 2.0 Pro Preview (medium) generates tokens at 115 tok/s versus 76 tok/s — about 34% faster. On time-to-first-token, Nova 2.0 Pro Preview (medium) responds in 13020ms vs 184480ms, which matters most for chat-style UIs.
Provider. Anthropic and Amazon sell to overlapping but distinct developer audiences: Anthropic tends to ship frontier reasoning models with premium positioning, while Amazon often prices more aggressively. Your existing vendor relationships, billing, and SLA preferences may matter as much as the raw numbers above.
Workload cost. Workload scenarios (per million requests at 30M input + 15M output tokens): Claude Sonnet 5 (Adaptive Reasoning, Max Effort) costs $210.00 ($2520/year); Nova 2.0 Pro Preview (medium) costs $187.50 ($2250/year). At a smaller 5M-input/2M-output scale (single-developer tool or prototype): Claude Sonnet 5 (Adaptive Reasoning, Max Effort) ≈ $30.00/run, Nova 2.0 Pro Preview (medium) ≈ $26.25/run. At agent/realtime scale (200M input / 100M output per million requests): Claude Sonnet 5 (Adaptive Reasoning, Max Effort) ≈ $1400/run, Nova 2.0 Pro Preview (medium) ≈ $1250/run. Nova 2.0 Pro Preview (medium) becomes more attractive at higher volume — the absolute per-token pricing difference compounds when you ship at scale.
Recommendation. Both models have legitimate use cases — the right answer depends on whether you are optimizing for benchmark ceiling, latency, or unit cost. Start with the cheaper / faster model, evaluate against your specific task, and only switch if the upgrade shows a meaningful lift.
Head-to-head deltas
Data from Artificial Analysis API — 12 benchmarks
Nova 2.0 Pro Preview (medium) is cheaper overall. Its blended price (3:1 input/output ratio) is $3.44/M tokens vs $4.00/M for Claude Sonnet 5 (Adaptive Reasoning, Max Effort).
Claude Sonnet 5 (Adaptive Reasoning, Max Effort) wins 2 out of 12 benchmarks compared to 0 for Nova 2.0 Pro Preview (medium). See the detailed benchmark chart above for per-category results.
Nova 2.0 Pro Preview (medium) generates tokens faster at 115 tok/s vs 76 tok/s. However, Nova 2.0 Pro Preview (medium) has lower time-to-first-token (13.02s vs 184.48s).
Choose based on your priorities: Nova 2.0 Pro Preview (medium) for lower cost, Claude Sonnet 5 (Adaptive Reasoning, Max Effort) for stronger benchmark performance, and Nova 2.0 Pro Preview (medium) for faster generation. For latency-sensitive apps, check the TTFT comparison above.