o3 vs Qwen3.7 Max
Full profiles: o3 (OpenAI) and Qwen3.7 Max (Alibaba): every benchmark, price and rank.
Updated · Data from 18 sources
As of 19 Nov 2026, o3 and Qwen3.7 Max are neck and neck on quality (60.2 vs 59.8 on the BenchLeader Index, inside the ±5.4 margin), and they cost about the same at $3.50 vs $3.75 per 1M tokens blended. Choose o3 for agents & tools (70 vs 49), maths (71 vs 58), and human preference (61 vs 56) and fast first answers (5.5 s vs 14.1 s); choose Qwen3.7 Max for instruction following (78 vs 65), knowledge (61 vs 55), and reasoning (64 vs 61), where it also streams faster.
Key facts (as of 19 Nov 2026)
- o3 scores 60.2 on the BenchLeader Index and Qwen3.7 Max scores 59.8 (BenchLeader, 19 Nov 2026).
- o3 costs $3.50 per 1M tokens blended (list price, 19 Nov 2026).
- Qwen3.7 Max costs $3.75 per 1M tokens blended (list price, 19 Nov 2026).
- Qwen3.7 Max streams faster than o3 (Artificial Analysis, 7-day median, 19 Nov 2026).
- Of the 17 benchmarks both models have, Qwen3.7 Max leads 17 (BenchLeader, 19 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 | o3 | 60 vs 59 category score; Arena Coding 1460 vs 1524 | Medium |
| Agents & tool use | o3 | 70 vs 49 category score; Terminal-Bench Hard 37.1% vs 50.8% | High |
| Maths | o3 | 71 vs 58 category score; Arena Maths 1448 vs 1481 | High |
| Reasoning | Qwen3.7 Max | 64 vs 61 category score; Arena Hard Prompts 1494 vs 1441 | Medium |
| Knowledge | Qwen3.7 Max | 61 vs 55 category score | Medium |
| Long documents | Qwen3.7 Max | 64 vs 62 category score; AA-LCR v1.1 79.0% vs 74.7% | Medium |
| Following instructions | Qwen3.7 Max | 78 vs 65 category score; IFBench 80.5% vs 71.4% | High |
| Chat quality (human preference) | o3 | 61 vs 56 on Arena votes | Medium |
| Lowest cost | Either | $3.50 vs $3.75 per 1M blended; agentic coding session $0.152 vs $0.180 | – |
| Fastest first answer | o3 | 5.5 s (o3) vs 14.1 s (Qwen3.7 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
- o3 and Qwen3.7 Max are neck and neck on quality (60.2 vs 59.8, within the ±5.4 margin).
- o3 is stronger in agents & tools, coding, human preference, maths.
- Qwen3.7 Max is stronger in instruction following, knowledge, long context, reasoning.
- They cost about the same ($3.50 per 1M blended).
At a glance
| Spec | o3 | Qwen3.7 Max |
|---|---|---|
| Developer | OpenAI | Alibaba |
| Released | 16 Apr 2025 | 19 May 2026 |
| Context window (tokens) | 200,000 | 1,000,000 |
| Compared at | o3 | Qwen3.7 Max |
| Open weights | No | No |
| Price in / out / cached input per 1M | $2.00 / $8.00 / $0.500 | $2.50 / $7.50 / $0.500 |
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 | o3 | Qwen3.7 Max | Leader |
|---|---|---|---|
| Reasoning | 60.9 | 64.2 | Qwen3.7 Max |
| Coding | 59.8 | 58.6 | o3 |
| Agents & tools | 70.3 | 48.6 | o3 |
| Maths | 70.6 | 57.5 | o3 |
| Knowledge | 54.9 | 60.9 | Qwen3.7 Max |
| Instruction following | 65.2 | 78.1 | Qwen3.7 Max |
| Human preference | 61.0 | 55.6 | o3 |
| Multimodal | 52.7 | – | not comparable |
| Long context | 61.9 | 64.1 | Qwen3.7 Max |
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: Multimodal.
Where sources disagree
Across the 17 benchmarks both have, o3 leads 0, Qwen3.7 Max leads 17. Every overall index that rates both puts Qwen3.7 Max ahead.
| Overall index | o3 | Qwen3.7 Max | Leader |
|---|---|---|---|
| BenchLeader Index | 60.2 | 59.8 | neck and neck |
| Arena Text | 1432 | 1474 | Qwen3.7 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 | o3 | Qwen3.7 Max |
|---|---|---|
| Blended price per 1M tokens | $3.50 | $3.75 |
| Input / output per 1M | $2.00 / $8.00 | $2.50 / $7.50 |
| Cached input per 1M | $0.500 | $0.500 |
| Price per task: chat reply | $0.0032 | $0.0033 |
| Price per task: summarize a 30-page report | $0.029 | $0.035 |
| Price per task: code edit | $0.024 | $0.026 |
| Price per task: agentic coding session | $0.152 | $0.180 |
| Price per task: structured extraction | $0.0056 | $0.0065 |
| Agentic coding session, input mostly cached | $0.085 | $0.090 |
| Output speed | 114 tok/s | 145 tok/s |
| Time to first answer | 5.5 s | 14.1 s |
| Time to a complete answer, last measured 13 Nov 2026 | 17.3 s | 31.2 s |
Each model is compared at its best-ranked reasoning-effort setting. o3 scores 60.2 and takes 5 s to its first answer token; o3 (high) scores 55.0 in 5 s, o3 (medium) scores 53.3 in 5 s. Qwen3.7 Max scores 59.8 and takes 14 s to its first answer token. 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
134 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 | o3 | Qwen3.7 Max |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 60.2 | 59.8 |
| Agents & tools score | BenchLeader | 70.3 | 48.6 |
| Coding score | BenchLeader | 59.8 | 58.6 |
| Human preference score | BenchLeader | 61.0 | 55.6 |
| Instruction following score | BenchLeader | 65.2 | 78.1 |
| Knowledge score | BenchLeader | 54.9 | 60.9 |
| Long context score | BenchLeader | 61.9 | 64.1 |
| Maths score | BenchLeader | 70.6 | 57.5 |
| Multimodal score | BenchLeader | 52.7 | – |
| Reasoning score | BenchLeader | 60.9 | 64.2 |
| Blended price $/M | BenchLeader | $3.50 | $3.75 |
| Output speed | BenchLeader | 114 tok/s | 145 tok/s |
| Time to first answer | BenchLeader | 5.5 s | 14.1 s |
| Context window | BenchLeader | 200,000 tokens | 1,000,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 81.8%† | 90.9% |
| Humanity's Last Exam | Scale AI / CAIS | 20.3%† | – |
| SimpleBench | SimpleBench | 53.1%† | 70.4% |
| Arena Hard Prompts | Arena (LMArena) | 1441 | 1494 |
| LiveBench 1.0 Reasoning | LiveBench | – | 83.3% |
| GPQA Diamond (AA) | Artificial Analysis | 82.7% | 92.3% |
| Humanity's Last Exam (AA) | Artificial Analysis | 20.1% | 40.5% |
| GPQA Diamond (Vals) | Vals AI | 84.1%† | 90.2% |
| Kagi LLM Benchmark | Kagi LLM Benchmark | 67.6% | – |
| GPQA Diamond (HELM) | HELM Capabilities | 75.3% | – |
| ARC-AGI-1 | ARC Prize | 60.8%† | – |
| ARC-AGI-2 | ARC Prize | 6.5%† | – |
| CritPt | Artificial Analysis | 1.1% | 13.4% |
| SciPredict | Scale AI SEAL | 17.9% | – |
| Arena Expert | Arena (LMArena) | 1448 | 1508 |
| Chess Puzzles | Epoch AI Benchmarking Hub | 36.0%† | 27.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 29.0%† | 32.0% |
| CL-bench | Epoch AI Benchmarking Hub | 17.8%† | – |
| LMCA | Epoch AI Benchmarking Hub | 39.7%† | 44.0% |
| DTBench | Epoch AI Benchmarking Hub | 84.8%† | 92.3% |
| ForecastBench | ForecastBench | 64.2% | – |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | 62.3%† | 77.3% |
| SciCode | SciCode | – | 49.5% |
| WeirdML | WeirdML | 52.4%† | – |
| GSO-Bench | GSO-Bench | 8.8%† | – |
Show all 134 benchmarks
| Metric | Publisher | o3 | Qwen3.7 Max |
|---|---|---|---|
| Coding | |||
| Arena Coding | Arena (LMArena) | 1460 | 1524 |
| Arena WebDev | Arena (LMArena) | – | 1516 |
| LiveBench 1.0 Coding | LiveBench | – | 74.2% |
| LiveCodeBench | Vals AI | 83.9%† | 87.1% |
| SWE-bench (Vals) | Vals AI | – | 68.8% |
| Aider Polyglot | Aider polyglot leaderboard | 76.9% | – |
| SWE-bench Verified (bash only) | SWE-bench | 58.4% | – |
| SWE-bench Verified (any scaffold) | SWE-bench | 58.4% | – |
| Code Migration | Vals AI | – | 13.1% |
| Vibe Code Bench v1.1 | Vals AI | – | 47.7% |
| CadEval | Epoch AI Benchmarking Hub | 74.0%† | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 933.5† | 1189.4 |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | – | 1.5% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 74.5% |
| Startrise Frontend | Startrise | – | 63.6% |
| GameBench 2 Overall | DuelLab GameBench 2 | 34.1 | – |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 30.5† | – |
| LiveCodeBench (AA) | Artificial Analysis | 80.8% | – |
| Agents & tools | |||
| GDPval | OpenAI | 30.8%† | – |
| Arena Agent | Arena (LMArena) | – | -8.3 |
| LiveBench Agentic Coding | LiveBench | – | 43.6% |
| Terminal-Bench Hard | Artificial Analysis | 37.1% | 50.8% |
| τ²-Bench Telecom (AA) | Artificial Analysis | 80.7% | 94.7% |
| Terminal-Bench 2.1 (Vals) | Vals AI | – | 61.0% |
| BFCL Overall | Berkeley Function Calling Leaderboard | 63.0% | – |
| GDPval-AA v2.1 | Artificial Analysis | – | 1134 |
| τ³-Banking (AA) | Artificial Analysis | – | 11.8% |
| ITBench SRE (AA) | Artificial Analysis | – | 42.5% |
| Analyst Agent (AA) | Artificial Analysis | – | 18.8% |
| Finance Agent v2 | Vals AI | – | 47.8% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 1.7% |
| Legal Research Bench | Vals AI | – | 25.5% |
| Tax Agent Bench | Vals AI | – | 15.8% |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 59.2% |
| Arena Search | Arena (LMArena) | 1197 | – |
| Arena Agent Task Completion | Arena (LMArena) | – | -10.6 |
| DeepResearch Bench | DeepResearch Bench | 45.2%† | – |
| METR Time Horizons | METR | 63.6% | – |
| GBAEval | Epoch AI Benchmarking Hub | – | 0.4% |
| AutomationBench | Artificial Analysis | – | 23.0% |
| GDP.pdf (AA) | Artificial Analysis | – | 8.6% |
| EnterpriseOps-Gym | Artificial Analysis | – | 45.0% |
| AA-Briefcase v1.1 | Artificial Analysis | – | 917 |
| Design Arena Agentic | Design Arena | – | 1171 |
| BrowseComp-Plus | BrowseComp-Plus | 50.5% | – |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 33.3%† | 64.6% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | – | 34.1% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 83.9%† | 95.6% |
| MATH Level 5 | Epoch AI Benchmarking Hub | 97.8%† | – |
| ProofBench | Vals AI | – | 26.0% |
| LiveBench Mathematics | LiveBench | – | 85.3% |
| AIME (Vals) | Vals AI | 85.3%† | – |
| MGSM | Vals AI | 91.8%† | – |
| IMO 2025 | MathArena | 16.7%† | – |
| Omni-MATH (HELM) | HELM Capabilities | 71.4% | – |
| MATH 500 | Vals AI | 94.6%† | – |
| Arena Maths | Arena (LMArena) | 1448 | 1481 |
| AIME 2025 (AA) | Artificial Analysis | 88.3% | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 49.4%† | 55.8% |
| LiveBench 1.0 Data Analysis | LiveBench | – | 71.8% |
| MMLU-Pro | Vals AI | 85.6%† | 89.3% |
| LegalBench | Vals AI | 83.8%† | 84.9% |
| CorpFin | Vals AI | 59.7%† | 63.7% |
| TaxEval | Vals AI | 74.6%† | 75.3% |
| MedQA | Vals AI | 96.1%† | – |
| PRBench Finance | Scale AI SEAL | 47.7% | – |
| PRBench Legal | Scale AI SEAL | 48.6% | – |
| MultiNRC | Scale AI SEAL | 45.5%† | – |
| MMLU-Pro (HELM) | HELM Capabilities | 85.9% | – |
| Excel Modeling Benchmark | Vals AI | – | 57.0% |
| MedCode | Vals AI | 47.3%† | 38.8% |
| MedScribe | Vals AI | 76.7%† | 79.4% |
| DrugDiscoveryBench | Scale AI SEAL | – | 29.3%† |
| EBR-bench | Epoch AI Benchmarking Hub | – | 9.5% |
| Instruction following | |||
| LiveBench Language | LiveBench | – | 79.7% |
| LiveBench Instruction Following | LiveBench | – | 74.0% |
| IFBench | Artificial Analysis | 71.4% | 80.5% |
| MultiChallenge | Scale AI SEAL | 56.6%† | – |
| TutorBench | Scale AI SEAL | 52.1%† | – |
| IFEval (HELM) | HELM Capabilities | 86.9% | – |
| WildBench (HELM) | HELM Capabilities | 86.1% | – |
| Arena Instruction Following | Arena (LMArena) | 1403 | 1466 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1432 | 1474 |
| EQ-Bench 4 | EQ-Bench | – | 1100 |
| EQ-Bench Creative Writing v3 | EQ-Bench | 1675 | – |
| Arena Creative Writing | Arena (LMArena) | 1383 | 1440 |
| Arena Multi-turn | Arena (LMArena) | 1420 | 1481 |
| Arena Longer Queries | Arena (LMArena) | 1411 | 1491 |
| Lech Mazur Creative Writing | Lech Mazur | 83.9%† | – |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1214 | – |
| MMMU-Pro | Artificial Analysis | 70.1% | – |
| VISTA | Scale AI SEAL | 50.1%† | – |
| MMMU-Pro (Vals) | Vals AI | 80.4%† | – |
| MortgageTax | Vals AI | 65.7%† | – |
| VTB | Scale AI SEAL | 13.7% | – |
| GeoBench | GeoBench | 60.0%† | – |
| VPCT | Epoch AI Benchmarking Hub | 52.0%† | – |
| Long context | |||
| Fiction.LiveBench 120k | Fiction.live | 100.0%† | – |
| AA-LCR v1.1 | Artificial Analysis | 74.7% | 79.0% |
| MLCR | Artificial Analysis | – | 14.4% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 147.1 | 154.6 |
| LiveBench 1.0 | LiveBench | – | 73.1% |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 20.2 | 29.5 |
| AA Progress Index v1 | Artificial Analysis | 258.3 | 293.8 |
| HELM Capabilities mean | HELM Capabilities | 81.1% | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 16.0% | – |
| MASK | Scale AI SEAL | 84.5%† | – |
| PropensityBench | Scale AI SEAL | 10.5% | – |
| Metadata | |||
| AA Openness Index | Artificial Analysis | 5.6 | – |
Data as of 2026-11-19. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261114-ca113f475f5e (https://benchmarks.duellab.org/releases/gb2-20261114-ca113f475f5e/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- Qwen3.7 Max: BenchLeader Index 60.3 → 59.8.
- Qwen3.7 Max: 1 score revision.
- o3: BenchLeader Index 61.2 → 60.4.
- o3: 1 score revision.
- o3: 1 new result (Arena Search).
- o3: 1 score revision.
- Qwen3.7 Max: BenchLeader Index 60.8 → 60.3.
- o3: 1 new result (AA Progress Index v1).
- o3: 1 score revision.
Related comparisons
- o3 vs Claude Opus 5.5
- o3 vs GPT-6 Astra
- o3 vs GPT-6.1 Sol
- o3 vs GPT-5.6 Sol
- o3 vs Gemini 4 Argon
- o3 vs Muse Wonder
- o3 vs Claude Sonnet 5.5
- o3 vs GPT-5.3-Codex
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
o3 vs Qwen3.7 Max: questions
- Is o3 better than Qwen3.7 Max?
- o3 and Qwen3.7 Max are neck and neck on quality (60.2 vs 59.8, within the ±5.4 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 19 Nov 2026, so check the category scores for your use.
- o3 vs Qwen3.7 Max: which is cheaper, and will that change?
- o3 is cheaper today: $3.50 against $3.75 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is o3 or Qwen3.7 Max better for coding?
- o3 scores higher in coding (60 vs 59 on the category index, where 50 is average). On Arena Coding, the top shared coding benchmark, it is 1460 for o3 against 1524 for Qwen3.7 Max.
- Which is faster, o3 or Qwen3.7 Max?
- Qwen3.7 Max streams faster (145 vs 114 output tokens per second, Artificial Analysis 7-day median).
- Is o3 or Qwen3.7 Max better for agents and tool use?
- o3 scores higher in agentic tasks (70 vs 49 on the category index, where 50 is average). On Terminal-Bench Hard, the top shared agents & tools benchmark, it is 37.1% for o3 against 50.8% for Qwen3.7 Max.
More questions (4)
- Which is better for long documents?
- Qwen3.7 Max scores higher on long-context benchmarks (64 vs 62). Qwen3.7 Max accepts more context: 1,000,000 against 200,000 tokens for o3.
- Which has lower latency?
- o3 answers first: 5.5 s against 14.1 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 o3 vs Qwen3.7 Max?
- Across the 17 benchmarks both have, o3 leads 0, Qwen3.7 Max leads 17. Every overall index that rates both puts Qwen3.7 Max ahead.
- When was Qwen3.7 Max released?
- Qwen3.7 Max was released on 19 May 2026.