GPT-5.6 Luna vs Qwen3.8 Max
Full profiles: GPT-5.6 Luna (OpenAI) and Qwen3.8 Max (Alibaba): every benchmark, price and rank.
Updated · Data from 16 sources
As of 16 Nov 2026, GPT-5.6 Luna and Qwen3.8 Max are neck and neck on quality (58.2 vs 58.7 on the BenchLeader Index, inside the ±5.2 margin), and GPT-5.6 Luna is 6.7× cheaper at $0.450 vs $3.00 per 1M tokens blended. Qwen3.8 Max leads in every category we measure both in. GPT-5.6 Luna streams ~1.5× faster.
Key facts (as of 16 Nov 2026)
- GPT-5.6 Luna scores 58.2 on the BenchLeader Index and Qwen3.8 Max scores 58.7 (BenchLeader, 16 Nov 2026).
- GPT-5.6 Luna costs $0.450 per 1M tokens blended (list price, 16 Nov 2026).
- Qwen3.8 Max costs $3.00 per 1M tokens blended (list price, 16 Nov 2026).
- GPT-5.6 Luna streams ~1.5× faster than Qwen3.8 Max (Artificial Analysis, 7-day median, 16 Nov 2026).
- Of the 25 benchmarks both models have, Qwen3.8 Max leads 22 (BenchLeader, 16 Nov 2026).
- On Arena WebDev, GPT-5.6 Luna is rated 1519 and Qwen3.8 Max 1676 (LMArena, as published by 16 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Qwen3.8 Max | 62 vs 56 category score; Arena WebDev 1676 vs 1519 | Medium |
| Agents & tool use | Qwen3.8 Max | 59 vs 57 category score; τ³-Banking (AA) 51.3% vs 28.7% | Medium |
| Maths | Either | 66 vs 65 category score; Arena Maths 1476 vs 1505 | Either |
| Reasoning | Either | 52 vs 52 category score; Arena Hard Prompts 1475 vs 1507 | Either |
| Long documents | Either | 66 vs 65 category score; AA-LCR v1.1 81.7% vs 80.3% | Either |
| Images and other media | Qwen3.8 Max | 66 vs 60 category score; Arena Vision 1312 vs 1256 | Medium |
| Chat quality (human preference) | Qwen3.8 Max | 68 vs 64 on Arena votes | Medium |
| Lowest cost | GPT-5.6 Luna | $0.450 vs $3.00 per 1M blended; agentic coding session $0.017 vs $0.144 | – |
| Fastest first answer | GPT-5.6 Luna | 28.1 s (GPT-5.6 Luna (xhigh)) vs 119.6 s (Qwen3.8 Max (max)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- GPT-5.6 Luna (xhigh) and Qwen3.8 Max (max) are neck and neck on quality (58.2 vs 58.7, within the ±5.2 margin).
- GPT-5.6 Luna (xhigh) is stronger in long context, maths, reasoning.
- Qwen3.8 Max (max) is stronger in agents & tools, coding, human preference, multimodal.
- GPT-5.6 Luna (xhigh) is 6.7× cheaper ($0.450 vs $3.00 per 1M blended).
- GPT-5.6 Luna (xhigh) streams 1.9× faster (110 vs 59 tokens per second).
At a glance
| Spec | GPT-5.6 Luna | Qwen3.8 Max |
|---|---|---|
| Developer | OpenAI | Alibaba |
| Released | 9 Jul 2026 | 3 Aug 2026 |
| Context window (tokens) | 1,050,000 | 1,000,000 |
| Compared at | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $0.200 / $1.20 / $0.020 | $2.00 / $6.00 / $0.250 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) | Leader |
|---|---|---|---|
| Reasoning | 51.8 | 51.5 | level |
| Coding | 56.4 | 61.6 | Qwen3.8 Max |
| Agents & tools | 57.1 | 58.5 | Qwen3.8 Max |
| Maths | 65.7 | 65.2 | level |
| Knowledge | – | 58.0 | not comparable |
| Human preference | 63.7 | 67.5 | Qwen3.8 Max |
| Multimodal | 59.9 | 65.7 | Qwen3.8 Max |
| Long context | 65.5 | 64.8 | level |
| Composite | – | 45.0 | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Knowledge, Instruction following.
Where sources disagree
Across the 25 benchmarks both have, GPT-5.6 Luna leads 3, Qwen3.8 Max leads 22. Every overall index that rates both puts Qwen3.8 Max ahead.
| Overall index | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 58.2 | 58.7 | neck and neck |
| AA Intelligence Index v5.0.1 | 23.8† | 33.6 | Qwen3.8 Max |
| Arena Text | 1453 | 1484 | Qwen3.8 Max |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) |
|---|---|---|
| Blended price per 1M tokens | $0.450 | $3.00 |
| Input / output per 1M | $0.200 / $1.20 | $2.00 / $6.00 |
| Cached input per 1M | $0.020 | $0.250 |
| Price per task: chat reply | $0.0004 | $0.0026 |
| Price per task: summarize a 30-page report | $0.0031 | $0.028 |
| Price per task: code edit | $0.0030 | $0.021 |
| Price per task: agentic coding session | $0.017 | $0.144 |
| Price per task: structured extraction | $0.0006 | $0.0052 |
| Agentic coding session, input mostly cached | $0.0087 | $0.065 |
| Cost per benchmark task (AA Index) | – | $10.02 |
| Output tokens per task (AA Index) | – | 171k |
| Of which reasoning | – | 120k |
| Time per task (AA Index) | – | 48 min |
| Output speed | 110 tok/s | 59 tok/s |
| Time to first answer | 28.1 s | 119.6 s |
| Time to a complete answer, last measured 13 Nov 2026 | 101.7 s | 294.1 s |
Each model is compared at its best-ranked reasoning-effort setting. GPT-5.6 Luna (xhigh) scores 58.2 and takes 28 s to its first answer token; GPT-5.6 Luna (max) scores 56.2 in 59 s, GPT-5.6 Luna (high) scores 54.7 in 11 s. Qwen3.8 Max (max) scores 58.7 and takes 120 s to its first answer token; Qwen3.8 Max (xhigh) scores 56.6 in 3 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
155 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 58.2 | 58.7 |
| Agents & tools score | BenchLeader | 57.1 | 58.5 |
| Coding score | BenchLeader | 56.4 | 61.6 |
| Human preference score | BenchLeader | 63.7 | 67.5 |
| Long context score | BenchLeader | 65.5 | 64.8 |
| Maths score | BenchLeader | 65.7 | 65.2 |
| Multimodal score | BenchLeader | 59.9 | 65.7 |
| Reasoning score | BenchLeader | 51.8 | 51.5 |
| Composite score | BenchLeader | – | 45.0 |
| Knowledge score | BenchLeader | – | 58.0 |
| Blended price $/M | BenchLeader | $0.450 | $3.00 |
| Output speed | BenchLeader | 110 tok/s | 59 tok/s |
| Time to first answer | BenchLeader | 28.1 s | 119.6 s |
| Context window | BenchLeader | 1,050,000 tokens | 1,000,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 91.6%† | 92.7%† |
| SimpleBench | SimpleBench | 46.8% | – |
| Arena Hard Prompts | Arena (LMArena) | 1475 | 1507 |
| LiveBench 2.0 Reasoning | LiveBench | – | 88.2% |
| LiveBench 1.0 Reasoning | LiveBench | 85.6%† | 88.2% |
| GPQA Diamond (AA) | Artificial Analysis | 89.5% | 92.8% |
| Humanity's Last Exam (AA) | Artificial Analysis | 37.0% | 43.1% |
| HLE-Diamond (AA) | Artificial Analysis | 20.9%† | 24.0% |
| GPQA Diamond (Vals) | Vals AI | 91.7%† | 93.7% |
| Kagi LLM Benchmark | Kagi LLM Benchmark | 49.1%† | – |
| ARC-AGI-1 | ARC Prize | 87.7% | – |
| ARC-AGI-2 | ARC Prize | 47.6% | – |
| ARC-AGI-3 | ARC Prize | 0.0% | – |
| CritPt | Artificial Analysis | 20.6% | 20.0% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 30.5%† | 29.5% |
| MysteryMechanism | Vals AI | 14.4%† | 23.9% |
| Arena Expert | Arena (LMArena) | 1490 | 1514 |
| Chess Puzzles | Epoch AI Benchmarking Hub | 43.0%† | 46.0%† |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 21.0%† | 38.0%† |
| LMCA | Epoch AI Benchmarking Hub | 48.5% | 46.2%† |
| DTBench | Epoch AI Benchmarking Hub | 89.1% | 92.0%† |
| Coding | |||
| SciCode | SciCode | 50.5% | 53.2% |
| WeirdML | WeirdML | 60.9%† | – |
| FrontierCode | Cognition | 39.8%† | – |
| Arena Coding | Arena (LMArena) | 1503 | 1523 |
Show all 155 benchmarks
| Metric | Publisher | GPT-5.6 Luna (xhigh) | Qwen3.8 Max (max) |
|---|---|---|---|
| Coding | |||
| Arena WebDev | Arena (LMArena) | 1519 | 1676 |
| LiveBench 2.0 Coding | LiveBench | – | 60.9% |
| LiveBench 1.0 Coding | LiveBench | 82.9%† | 72.9% |
| LiveCodeBench | Vals AI | – | 87.8% |
| IOI | Vals AI | 61.8%† | 68.9% |
| SWE-bench (Vals) | Vals AI | 93.0%† | 85.6% |
| Code Migration | Vals AI | 44.5%† | 24.0% |
| ProgramBench | Vals AI | 0.0%† | 0.0% |
| SWE-bench Multimodal | Vals AI | – | 42.5% |
| Vibe Code Bench 1-100 | Vals AI | 22.6%† | 12.8% |
| Vibe Code Bench v1.1 | Vals AI | 77.1%† | 64.7% |
| SWE Atlas v1.1 | Scale AI SEAL | – | 64.4% |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | – | 81.5% |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | – | 58.9% |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | – | 52.9% |
| MirrorCode | Epoch AI Benchmarking Hub | 21.1%† | 52.2%† |
| Surface Evolver Bench | Epoch AI Benchmarking Hub | 61.9%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 56.9% | 57.5%† |
| CursorBench | Cursor | 33.0% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 1667.4† | 1679.1† |
| FrontierSWE | Epoch AI Benchmarking Hub | – | 17.8%† |
| WeirdML v3 | WeirdML | 5.2% | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 3.5% | 38.9% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 77.9% | 88.8% |
| Startrise Frontend | Startrise | 65.0%† | 71.3% |
| GameBench 2 Overall | DuelLab GameBench 2 | 40.2† | 54.2† |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 54 | 69.6† |
| Design Arena Website | Design Arena | 1248† | 1269 |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 43.2% |
| FrontierCode 2.0 (AA) | Artificial Analysis | 17.6%† | 23.4% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 17.6%† | 27.0% |
| APEX-Agents | Mercor | 43.0%† | 63.3% |
| Arena Agent | Arena (LMArena) | – | 0.1 |
| LiveBench Long Agentic | LiveBench | – | 26.6% |
| LiveBench Browser | LiveBench | – | 49.4% |
| LiveBench Agentic Coding | LiveBench | 48.4%† | 64.7% |
| Terminal-Bench 2.1 (Vals) | Vals AI | 79.0%† | 67.4% |
| τ³-bench Banking | τ²-bench | – | 55.1%† |
| GDPval-AA v2.1, normalised score | Artificial Analysis | – | 58.3% |
| GDPval-AA v2.1 | Artificial Analysis | 1407 | 1674 |
| τ³-Banking (AA) | Artificial Analysis | 28.7% | 51.3% |
| ITBench SRE (AA) | Artificial Analysis | 40.3%† | 40.3% |
| Analyst Agent (AA) | Artificial Analysis | – | 45.0% |
| APEX-Agents (AA) | Artificial Analysis | 35.8%† | 42.4% |
| CyberBench | Vals AI | 73.6%† | 28.6% |
| Finance Agent v2 | Vals AI | 55.0%† | 50.6% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 1.3%† | 10.4% |
| Legal Research Bench | Vals AI | 36.5%† | 47.6% |
| SkillsBench | Vals AI | 60.5%† | 42.0% |
| SREBench | Vals AI | 1.5%† | – |
| Tax Agent Bench | Vals AI | 24.8%† | 32.1% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 11.6%† | 34.3% |
| Terminal-Bench Science | Vals AI | 0.0%† | 12.9% |
| Arena Search | Arena (LMArena) | – | 1227 |
| Arena Agent Task Completion | Arena (LMArena) | – | 4.5 |
| BALROG | BALROG | 45.6%† | – |
| Vending-Bench 2 | Andon Labs | 4094.7† | – |
| AutomationBench | Artificial Analysis | 42.6% | 56.2% |
| GDP.pdf (AA) | Artificial Analysis | 23.8% | 22.8% |
| EnterpriseOps-Gym | Artificial Analysis | 40.8%† | 47.6% |
| AA-Briefcase v1.1 | Artificial Analysis | 1267 | 1604 |
| GDPx | GDPx by SemiAnalysis | – | 3.5% |
| Design Arena Agentic | Design Arena | 1149† | 1264 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 3.3%† | 11.9% |
| MKA (AA) | Artificial Analysis | – | 37.4% |
| FACTS Search | 87.0%† | 85.0% | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 82.1%† | 74.7%† |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 61.0%† | 46.3%† |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 98.3%† | 100.0%† |
| ProofBench | Vals AI | 60.0%† | 58.0% |
| LiveBench Mathematics | LiveBench | 87.2%† | 91.3% |
| Arena Maths | Arena (LMArena) | 1476 | 1505 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 41.0%† | 47.3%† |
| LiveBench 2.0 Data Analysis | LiveBench | – | 78.4% |
| LiveBench Research | LiveBench | – | 72.8% |
| LiveBench 1.0 Data Analysis | LiveBench | 78.0%† | 78.4% |
| AA-Omniscience v1.1 | Artificial Analysis | -17.1† | 10.5 |
| MMLU-Pro | Vals AI | 86.0%† | 88.6% |
| LegalBench | Vals AI | 84.0%† | 83.6% |
| CorpFin | Vals AI | 64.2%† | 65.8% |
| TaxEval | Vals AI | 76.2%† | 75.5% |
| BioMysteryBench | Vals AI | 61.5%† | – |
| Excel Modeling Benchmark | Vals AI | 67.1%† | 60.1% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 66.0% |
| MedCode | Vals AI | 42.4%† | 40.7% |
| MedScribe | Vals AI | 84.4%† | 85.0% |
| Public Benefits Bench v1.1 | Vals AI | 61.2%† | 67.1% |
| Access to Justice Benchmark | Vals AI | 65.6%† | 76.4% |
| Financial Coaching Bench | Vals AI | 69.3%† | 75.6% |
| Immigration Bench | Vals AI | – | 70.1% |
| GDP.pdf | Epoch AI Benchmarking Hub | 22.7%† | 23.2%† |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 39.1%† | 29.5% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 7.7%† | 72.9% |
| FACTS Parametric | 50.6%† | 30.8% | |
| Instruction following | |||
| LiveBench Language | LiveBench | 72.6%† | 79.7% |
| LiveBench Instruction Following | LiveBench | 60.1%† | 74.1% |
| Arena Instruction Following | Arena (LMArena) | 1443 | 1478 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1453 | 1484 |
| EQ-Bench 4 | EQ-Bench | 1146† | – |
| EQ-Bench Creative Writing v3 | EQ-Bench | 1832† | – |
| Arena Creative Writing | Arena (LMArena) | 1411 | 1472 |
| Arena Multi-turn | Arena (LMArena) | 1458 | 1495 |
| Arena Longer Queries | Arena (LMArena) | 1452 | 1496 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1256 | 1312 |
| LiveBench Design | LiveBench | – | 76.6% |
| MMMU-Pro | Artificial Analysis | 78.5% | 82.8% |
| MMMU-Pro (Vals) | Vals AI | 85.0%† | 88.0% |
| MortgageTax | Vals AI | 67.3%† | 64.0% |
| SAGE | Vals AI | – | 81.8% |
| Humanity's Sixth Sense | Scale AI SEAL | 21.6%† | 33.0% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 22.6%† | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 42.5%† | 20.0%† |
| VIALS-AA | Handshake AI | – | 31.7% |
| Code2Video-AA | Artificial Analysis | – | 1021 |
| AA-3D | Artificial Analysis | – | 954 |
| FACTS Multimodal | 37.4%† | 51.5% | |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 81.7% | 80.3% |
| Arena Document | Arena (LMArena) | 1460 | – |
| MLCR | Artificial Analysis | 15.0% | 26.1% |
| FACTS Grounding | 52.5%† | 73.8% | |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 156.7† | 157.6 |
| LiveBench 2.0 | LiveBench | – | 60.9% |
| LiveBench 1.0 | LiveBench | 73.6%† | 78.5% |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 23.8† | 33.6 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 34.6 | 45.4 |
| AA Progress Index v1 | Artificial Analysis | 305.8 | 323.7 |
| Vals RSI Index | Vals AI | – | 21.8% |
| Vals Science Index | Vals AI | 25.3%† | – |
| Vals Multimodal Index | Vals AI | 69.1%† | 65.4% |
| Safety & honesty | |||
| DistressBench | Scale AI SEAL | 72.7%† | 77.6% |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- Qwen3.8 Max (max): 1 score revision.
- GPT-5.6 Luna (xhigh): 2 new results (GDPval-AA v2.1, AA-Briefcase v1.1).
- Qwen3.8 Max (max): 1 score revision.
- GPT-5.6 Luna (xhigh): 1 score revision.
- Qwen3.8 Max (max): 6 score revisions.
- Qwen3.8 Max (max): 1 score revision.
- Qwen3.8 Max (max): 1 new result (FACTS Search); 1 score revision.
- GPT-5.6 Luna (xhigh): 1 score revision.
- Qwen3.8 Max (max): 1 score revision; cached input price moved from $0.22/M to $0.25/M (13% up).
- Qwen3.8 Max (max): cached input price moved from $0.25/M to $0.22/M (11% down).
Related comparisons
- Qwen3.8 Max vs GPT-6 Astra
- GPT-5.5 vs Qwen3.8 Max
- GPT-5.6 Luna vs GPT-6.1 Sol
- GPT-5.6 Luna vs Gemini 4 Argon
- GPT-5.6 Luna vs Claude Sonnet 5.5
- GPT-5.6 Luna vs GPT-5.5
- GPT-5.6 Luna vs Inkling 2 Small
- Qwen3.8 Max vs Muse Spark
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
GPT-5.6 Luna vs Qwen3.8 Max: questions
- Is GPT-5.6 Luna better than Qwen3.8 Max?
- GPT-5.6 Luna (xhigh) and Qwen3.8 Max (max) are neck and neck on quality (58.2 vs 58.7, within the ±5.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
- GPT-5.6 Luna vs Qwen3.8 Max: which is cheaper, and will that change?
- GPT-5.6 Luna is cheaper today: $0.450 against $3.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is GPT-5.6 Luna or Qwen3.8 Max better for coding?
- Qwen3.8 Max scores higher in coding (62 vs 56 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1519 for GPT-5.6 Luna against 1676 for Qwen3.8 Max.
- Which is faster, GPT-5.6 Luna or Qwen3.8 Max?
- GPT-5.6 Luna streams ~1.5× faster (110 vs 59 output tokens per second, Artificial Analysis 7-day median).
- Is GPT-5.6 Luna or Qwen3.8 Max better for agents and tool use?
- Qwen3.8 Max scores higher in agentic tasks (59 vs 57 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 28.7% for GPT-5.6 Luna against 51.3% for Qwen3.8 Max.
More questions (4)
- Which is better for long documents?
- They are level on long-context benchmarks (66 vs 65). Both take about 1.1M tokens: GPT-5.6 Luna 1,050,000, Qwen3.8 Max 1,000,000.
- Which has lower latency?
- GPT-5.6 Luna answers first: 28.1 s against 119.6 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on GPT-5.6 Luna vs Qwen3.8 Max?
- Across the 25 benchmarks both have, GPT-5.6 Luna leads 3, Qwen3.8 Max leads 22. Every overall index that rates both puts Qwen3.8 Max ahead.
- When was Qwen3.8 Max released?
- Qwen3.8 Max was released on 3 Aug 2026.