STEM thinking duel: Claude 3.7 Sonnet ($3/$15) vs OpenAI o3-mini ($1.10/$4.40).
claude-3-7-sonnet-20250219 · anthropic
o3-mini · openai
Benchmark
| Workload Scenario | Claude 3.7 Sonnet (Hybrid Reasoning) | o3-mini | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $3.00 | $1.10 | +$1.90 |
| 1M output tokens (generation) | $15.00 | $4.40 | +$10.60 |
| 1M tokens · 70% input / 30% output mix | $6.60 | $2.09 | +$4.51 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.0186 | $0.00682 | +$0.0118 |
| Monthly scale (10K requests / day) | $5,580.00 | $2,046.00 | +$3,534.00 |
Negative difference = Claude 3.7 Sonnet (Hybrid Reasoning) is cheaper. Positive = o3-mini is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
o3-mini is cheaper on both input and output rates, so it costs less at every input/output mix. Price alone still isn't the whole decision: capability, latency and context limits (Claude 3.7 Sonnet (Hybrid Reasoning): 200K, o3-mini: 200K) may justify the premium for your task.
FAQ
On input, o3-mini is cheaper ($1.10/M vs $3.00/M). On output, o3-mini is cheaper ($4.40/M vs $15.00/M). The same model is cheaper on both sides, so it wins at every mix.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.0186 on Claude 3.7 Sonnet (Hybrid Reasoning) and $0.00682 on o3-mini — o3-mini is 2.7× more expensive for that workload.
Claude 3.7 Sonnet (Hybrid Reasoning) supports 200,000 tokens (64K max output); o3-mini supports 200,000 (100K max output). They are equal.