Dedicated reasoning model versus the general tier that replaced it for most tasks.
o3 · openai
gpt-5.6-terra · openai
Benchmark
| Workload Scenario | o3 | GPT-5.6 Terra | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $2.00 | $2.00 | — |
| 1M output tokens (generation) | $8.00 | $12.00 | −$4.00 |
| 1M tokens · 70% input / 30% output mix | $3.80 | $5.00 | −$1.20 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.0114 | $0.014 | −$0.0026 |
| Monthly scale (10K requests / day) | $3,420.00 | $4,200.00 | −$780.00 |
Negative difference = o3 is cheaper. Positive = GPT-5.6 Terra is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
GPT-5.6 Terra has the cheaper input rate, while o3 has the cheaper output rate. The crossover happens when output makes up about 0% of your total tokens.
Below that share (retrieval, summarization, extraction — lots of context in, little text out) GPT-5.6 Terra is cheaper. Above it (generation, translation, coding — long completions) o3 wins.
FAQ
On input, GPT-5.6 Terra is cheaper ($2.00/M vs $2.00/M). On output, o3 is cheaper ($8.00/M vs $12.00/M). For workloads where more than 0% of tokens are output, the output-cheaper model wins overall.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.0114 on o3 and $0.014 on GPT-5.6 Terra — o3 is 1.2× cheaper for that workload.
o3 supports 200,000 tokens (100K max output); GPT-5.6 Terra supports 1,050,000 (128K max output). GPT-5.6 Terra fits 5.3× more context, which matters for long documents and agents.