Doubao Lite 1.5 vs GPT-5.6 Terra benchmark

Side-by-Side Benchmark Matrix

Recommended Workload Routing

Frequently Asked Questions

Which is better overall, Doubao Lite 1.5 or GPT-5.6 Terra?

GPT-5.6 Terra has the higher preference Elo in our latest snapshot (1,548). “Better” still depends on coding, price, and latency — see the table.

Which is better at coding, Doubao Lite 1.5 or GPT-5.6 Terra?

GPT-5.6 Terra leads SWE-bench at 70.4% vs 46%. SWE-bench is one harness, not your repo.

Which is cheaper to run in production?

Doubao Lite 1.5 output tokens are $0.1/1M versus $12/1M. Input prices and retry rates still move the real bill.

Which is faster, Doubao Lite 1.5 or GPT-5.6 Terra?

Doubao Lite 1.5 has the lower time-to-first-token (110 ms vs 160 ms). Tokens/sec is a separate column if you care about long completions.

Which has the larger context window?

GPT-5.6 Terra accepts 1M tokens versus 33k.

Can I self-host either model?

Neither model is marked open-weights here. You are comparing hosted APIs.

Are these scores from CompareLLM’s own evals?

No. V1 aggregates public leaderboards (and optional first-party latency pings). Each cell names its source and date. Read /methodology.

What does preference Elo mean on this page?

Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent. Doubao Lite 1.5 and GPT-5.6 Terra Elo cells are dated snapshots from the named source — we do not compute Arena ourselves.