GPT-5.6 Luna vs Qwen3.8-Max.
Verdict by category, tale of the tape, shared benchmarks only, and cost at three workloads — all from the tracked data.
Verdict by category
| Category | GPT-5.6 Luna | Qwen3.8-Max | Winner |
|---|---|---|---|
| League Score | 66.7 | 65.0 | GPT-5.6 Luna |
| Coding (SWE-bench %) | — | — | — |
| Reasoning (GPQA %) | 92.3 | 92.6 | Qwen3.8-Max |
| Context (tokens) | 1050000 | 1000000 | GPT-5.6 Luna |
| Speed (tok/s) | — | 37 | — |
| Price (blended $/1M) | $0.45 | $3.00 | GPT-5.6 Luna (cheaper) |
Tale of the tape
| Spec | GPT-5.6 Luna | Qwen3.8-Max |
|---|---|---|
| Maker | OpenAI | Alibaba Qwen |
| Released | Jul 9, 2026 | Aug 3, 2026 |
| League Score | 67 | 65 |
| Context | 1,050,000 tokens | 1,000,000 tokens |
| Max output | 128,000 | 131,072 |
| $/1M input | $0.20 | $2.00 |
| $/1M output | $1.20 | $6.00 |
| Weights | Closed | Open |
| Speed | — | 37 t/s |
Head-to-head benchmarks (shared only)
| Benchmark | GPT-5.6 Luna | Qwen3.8-Max |
|---|---|---|
| LiveBench overall | 66.1 | 61.6 |
| GPQA Diamond | 92.3 | 92.6 |
| MMLU-Pro | 86.0 | 88.6 |
Cost at three workloads
| Workload | GPT-5.6 Luna | Qwen3.8-Max |
|---|---|---|
| Light — 1M in / 250K out | $0.50 | $3.50 |
| Standard — 10M in / 2.5M out | $5.00 | $35.00 |
| Long-context — 50M in / 5M out | $16.00 | $130.00 |
The pick
Pick GPT-5.6 Luna if you want the higher overall League Score; you need the bigger context window; you want the lower running cost.
Pick Qwen3.8-Max if you want downloadable weights / self-hosting.
All figures from the model database — sources and dates on each model page. Cost your own workload or compare a different pair.
Frequently asked questions
Which is better overall, GPT-5.6 Luna or Qwen3.8-Max?
GPT-5.6 Luna leads on overall League Score, 67 to 65 — a gap of 2 points on a composite of sourced benchmarks. That is overall form, not every category: check the per-category verdicts before choosing.
Which is cheaper, GPT-5.6 Luna or Qwen3.8-Max?
GPT-5.6 Luna is cheaper: $0.45 vs $3.00 blended per 1M tokens (3:1 input/output mix) — about 6.7x less per token.
Which has the bigger context window, GPT-5.6 Luna or Qwen3.8-Max?
GPT-5.6 Luna, with 1.1M of context against 1M for Qwen3.8-Max.
Which is better at reasoning, GPT-5.6 Luna or Qwen3.8-Max?
Qwen3.8-Max leads on GPQA Diamond, 92.6% vs 92.3% on graduate-level science questions.
Are GPT-5.6 Luna and Qwen3.8-Max open source?
GPT-5.6 Luna is closed weights — API access only; Qwen3.8-Max is open weights — it can be downloaded and self-hosted.