MiMo-V2.6-Pro vs Qwen4 Max
Full profiles: MiMo-V2.6-Pro (Xiaomi) and Qwen4 Max (Alibaba): every benchmark, price and rank.
Updated · Data from 8 sources
As of 19 Nov 2026, there are not enough shared benchmarks for a verdict, and MiMo-V2.6-Pro is 4.1× cheaper at $0.544 vs $2.25 per 1M tokens blended. Choose MiMo-V2.6-Pro for coding (58 vs 55) and fast first answers (42.8 s vs 139.2 s); choose Qwen4 Max for knowledge (59 vs 43) and reasoning (59 vs 50).
Key facts (as of 19 Nov 2026)
- MiMo-V2.6-Pro costs $0.544 per 1M tokens blended (list price, 19 Nov 2026).
- Qwen4 Max costs $2.25 per 1M tokens blended (list price, 19 Nov 2026).
- Of the 15 benchmarks both models have, Qwen4 Max leads 15 (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 | MiMo-V2.6-Pro | 58 vs 55 category score; FrontierCode 2.0 (AA) 24.0% vs 26.4% | Medium |
| Reasoning | Qwen4 Max | 59 vs 50 category score; HLE-Diamond (AA) 44.9% vs 15.2% | High |
| Knowledge | Qwen4 Max | 59 vs 43 category score; AA-Omniscience v1.1 27.7 vs -4.5 | High |
| Chat quality (human preference) | – | Only MiMo-V2.6-Pro has an Arena rating (67 category score) | Not comparable |
| Lowest cost | MiMo-V2.6-Pro | $0.548 vs $2.25 per 1M blended; agentic coding session $0.030 vs $0.108 | – |
| Fastest first answer | MiMo-V2.6-Pro | 42.8 s (MiMo-V2.6-Pro) vs 139.2 s (Qwen4 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
- Not enough shared benchmarks for a verdict.
- MiMo-V2.6-Pro is stronger in coding.
- Qwen4 Max is stronger in knowledge, reasoning.
- MiMo-V2.6-Pro is 4.1× cheaper ($0.544 vs $2.25 per 1M blended).
At a glance
| Spec | MiMo-V2.6-Pro | Qwen4 Max |
|---|---|---|
| Developer | Xiaomi | Alibaba |
| Released | 21 Sep 2026 | 16 Nov 2026 |
| Context window (tokens) | 1,048,576 | 991,808 |
| Compared at | MiMo-V2.6-Pro | Qwen4 Max |
| Open weights | Yes | No |
| Price in / out / cached input per 1M | $0.440 / $0.870 / $0.0036 | $1.50 / $4.50 / $0.071 |
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 | MiMo-V2.6-Pro | Qwen4 Max | Leader |
|---|---|---|---|
| Reasoning | 49.5 | 58.5 | Qwen4 Max |
| Coding | 58.1 | 55.1 | MiMo-V2.6-Pro |
| Agents & tools | 53.1 | – | not comparable |
| Maths | 65.4 | – | not comparable |
| Knowledge | 42.6 | 59.0 | Qwen4 Max |
| Human preference | 67.0 | – | not comparable |
| Multimodal | 60.2 | – | not comparable |
| Long context | 67.9 | – | 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: Agents & tools, Maths, Instruction following, Long context, Multimodal, Human preference.
Where sources disagree
Across the 15 benchmarks both have, MiMo-V2.6-Pro leads 0, Qwen4 Max leads 15. Every overall index that rates both puts Qwen4 Max ahead.
| Overall index | MiMo-V2.6-Pro | Qwen4 Max | Leader |
|---|---|---|---|
| AA Intelligence Index v5.0.1 | 29.9 | 42.9 | Qwen4 Max |
| Vals Index | 57.8 | 63.2† | Qwen4 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 | MiMo-V2.6-Pro | Qwen4 Max |
|---|---|---|
| Blended price per 1M tokens | $0.548 | $2.25 |
| Input / output per 1M | $0.440 / $0.870 | $1.50 / $4.50 |
| Cached input per 1M | $0.0036 | $0.071 |
| Price per task: chat reply | $0.0004 | $0.0020 |
| Price per task: summarize a 30-page report | $0.0058 | $0.021 |
| Price per task: code edit | $0.0039 | $0.016 |
| Price per task: agentic coding session | $0.030 | $0.108 |
| Price per task: structured extraction | $0.0011 | $0.0039 |
| Agentic coding session, input mostly cached | $0.010 | $0.044 |
| Cost per benchmark task (AA Index) | $0.257 | $3.27 |
| Output tokens per task (AA Index) | 121k | 105k |
| Of which reasoning | 86k | 69k |
| Time per task (AA Index) | 36 min | 30 min |
| Output speed | 56 tok/s | 59 tok/s |
| Time to first answer | 42.8 s | 139.2 s |
| Time to a complete answer, last measured 13 Nov 2026 | 95.7 s | – |
Per token is not per task: an average task of Artificial Analysis' Intelligence Index cost $0.257 on MiMo-V2.6-Pro and $3.27 on Qwen4 Max, and MiMo-V2.6-Pro wrote 1.2× as many output tokens for the same work.
Each model is compared at its best-ranked reasoning-effort setting. MiMo-V2.6-Pro scores 56.9 and takes 43 s to its first answer token. Qwen4 Max and takes 139 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
95 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 | MiMo-V2.6-Pro | Qwen4 Max |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 56.9 | – |
| Agents & tools score | BenchLeader | 53.1 | – |
| Coding score | BenchLeader | 58.1 | 55.1 |
| Human preference score | BenchLeader | 67.0 | – |
| Knowledge score | BenchLeader | 42.6 | 59.0 |
| Long context score | BenchLeader | 67.9 | – |
| Maths score | BenchLeader | 65.4 | – |
| Multimodal score | BenchLeader | 60.2 | – |
| Reasoning score | BenchLeader | 49.5 | 58.5 |
| Blended price $/M | BenchLeader | $0.548 | $2.25 |
| Output speed | BenchLeader | 56 tok/s | 59 tok/s |
| Time to first answer | BenchLeader | 42.8 s | 139.2 s |
| Context window | BenchLeader | 1,048,576 tokens | 991,808 tokens |
| Reasoning | |||
| Arena Hard Prompts | Arena (LMArena) | 1504 | – |
| LiveBench 2.0 Reasoning | LiveBench | 79.6%† | 92.2%† |
| Humanity's Last Exam (AA) | Artificial Analysis | 49.4% | – |
| HLE-Diamond (AA) | Artificial Analysis | 15.2% | 44.9% |
| CritPt | Artificial Analysis | 26.6% | – |
| CritPt v2 hard subset (AA) | Artificial Analysis | 34.0% | 43.0% |
| MysteryMechanism | Vals AI | 15.3% | – |
| Arena Expert | Arena (LMArena) | 1526 | – |
| Coding | |||
| SciCode | SciCode | 60.9% | – |
| Arena Coding | Arena (LMArena) | 1538 | – |
| Arena WebDev | Arena (LMArena) | 1623 | 1750† |
| LiveBench 2.0 Coding | LiveBench | 64.5%† | 69.8%† |
| IOI | Vals AI | 39.3% | – |
| Code Migration | Vals AI | 43.0% | – |
| Code Migration v1.1 | Vals AI | 48.8% | – |
| ProgramBench | Vals AI | 0.5% | – |
| SWE-bench Multimodal | Vals AI | 41.3% | – |
| Vibe Code Bench v1.1 | Vals AI | 85.2% | 86.7%† |
| Vibe Code Bench 1-100 v1.1 | Vals AI | 26.4% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 1158.1 | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 34.9% | 48.5% |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 49.8† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 45.4% | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 24.0% | 26.4% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 27.0%† | – |
Show all 95 benchmarks
| Metric | Publisher | MiMo-V2.6-Pro | Qwen4 Max |
|---|---|---|---|
| Agents & tools | |||
| APEX-Agents | Mercor | 54.4% | – |
| Arena Agent | Arena (LMArena) | 1.4 | 7.3† |
| LiveBench Long Agentic | LiveBench | 36.4%† | 46.6%† |
| LiveBench Browser | LiveBench | 77.3%† | 80.7%† |
| Terminal-Bench 2.1 (Vals) | Vals AI | 67.8% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 59.1% | – |
| GDPval-AA v2.1 | Artificial Analysis | 1692 | 1809 |
| Analyst Agent (AA) | Artificial Analysis | 37.5% | 46.3% |
| APEX-Agents (AA) | Artificial Analysis | – | 59.1% |
| CyberBench | Vals AI | 72.9% | – |
| Finance Agent v2 | Vals AI | 57.3% | – |
| Finance Agent v2.1 | Vals AI | 64.3% | 66.7%† |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 13.8% | 21.3%† |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 10.8% | – |
| Legal Research Bench | Vals AI | 47.1% | 41.4%† |
| SREBench | Vals AI | 3.0% | – |
| Tax Agent Bench | Vals AI | 32.4% | 40.4%† |
| Terminal-Bench 4.0 (Vals) | Vals AI | 31.3% | 48.5%† |
| Terminal-Bench Science | Vals AI | 2.9% | – |
| Arena Agent Task Completion | Arena (LMArena) | 3.5 | 8.8† |
| AutomationBench | Artificial Analysis | 58.6% | 72.6% |
| GDP.pdf (AA) | Artificial Analysis | 19.2% | 26.2% |
| Harvey LAB-AA v1.1 | Harvey | – | 9.4% |
| EnterpriseOps-Gym | Artificial Analysis | – | 48.5% |
| AA-Briefcase v1.1 | Artificial Analysis | 1512 | 1686 |
| Design Arena Agentic | Design Arena | 1263 | 1356 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 5.7% | 23.3% |
| CWE-Bench (AA) | Artificial Analysis | 63.3% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 26.2% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 78.6% | – |
| MKA (AA) | Artificial Analysis | 37.3% | – |
| Maths | |||
| ProofBench | Vals AI | 70.0% | – |
| Arena Maths | Arena (LMArena) | 1482 | – |
| Knowledge | |||
| LiveBench 2.0 Data Analysis | LiveBench | 72.3%† | 83.4%† |
| LiveBench Research | LiveBench | 84.8%† | 84.7%† |
| AA-Omniscience v1.1 | Artificial Analysis | -4.5 | 27.7 |
| Excel Modeling Benchmark | Vals AI | 62.9% | – |
| Excel Modeling Benchmark v1.1 | Vals AI | 67.3% | 70.8%† |
| MedCode | Vals AI | 45.0% | – |
| MedScribe | Vals AI | 88.3% | – |
| Public Benefits Bench v1.1 | Vals AI | 68.9% | – |
| Access to Justice Benchmark | Vals AI | 70.0% | – |
| Financial Coaching Bench | Vals AI | 75.4% | – |
| Immigration Bench | Vals AI | 67.5% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 34.8% | 43.7% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 39.8% | 71.6% |
| Instruction following | |||
| Arena Instruction Following | Arena (LMArena) | 1473 | – |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1480 | – |
| Arena Creative Writing | Arena (LMArena) | 1449 | – |
| Arena Multi-turn | Arena (LMArena) | 1463 | – |
| Arena Longer Queries | Arena (LMArena) | 1497 | – |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1266 | – |
| LiveBench Design | LiveBench | 74.3%† | 79.1%† |
| SAGE | Vals AI | 84.5% | – |
| Humanity's Sixth Sense | Scale AI SEAL | 25.2%† | – |
| VIALS-AA | Handshake AI | 43.5% | – |
| Code2Video-AA | Artificial Analysis | 1052 | – |
| AA-3D | Artificial Analysis | 983 | – |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 86.3% | – |
| MLCR | Artificial Analysis | 12.8% | 19.4% |
| Composite | |||
| LiveBench 2.0 | LiveBench | 66.1%† | 73.1%† |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 29.9 | 42.9 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 46.3 | – |
| AA Progress Index v1 | Artificial Analysis | 321.2 | – |
| Vals Index | Vals AI | 57.8 | 63.2† |
| Vals Finance Index | Vals AI | 57.9% | 62.0%† |
| Vals Router Index | Vals AI | 57.8% | – |
| Vals Legal Index | Vals AI | 72.1% | 70.7%† |
| AA Cyber Index | Artificial Analysis | 56.1 | – |
| Metadata | |||
| AA Openness Index | Artificial Analysis | 52.8 | – |
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
- MiMo-V2.6-Pro: 2 score revisions.
- MiMo-V2.6-Pro: 1 score revision.
- MiMo-V2.6-Pro: 1 score revision.
- MiMo-V2.6-Pro: 6 score revisions.
- MiMo-V2.6-Pro: 1 score revision.
- MiMo-V2.6-Pro: 1 score revision.
- MiMo-V2.6-Pro: BenchLeader Index 61.1 → 57.2.
- MiMo-V2.6-Pro: 5 new results (AA Intelligence Index v5.0.1, AA Progress Index v1, HLE-Diamond (AA)…); 3 score revisions; MiMo-V2.6-Pro cost per task moved from $0.13/M to $0.26/M (93% up).
- MiMo-V2.6-Pro: BenchLeader Index 61.7 → 61.1.
- MiMo-V2.6-Pro: 3 score revisions.
Related comparisons
- MiMo-V2.6-Pro vs GPT-6 Astra
- MiMo-V2.6-Pro vs GPT-6.1 Sol
- MiMo-V2.6-Pro vs Claude Fable 5
- MiMo-V2.6-Pro vs GPT-5.6 Sol
- Qwen4 Max vs Qwen3.8 Max
- Qwen4 Max vs Inkling 2 Small
- MiMo-V2.6-Pro vs Qwen3.8 Max
- MiMo-V2.6-Pro vs Gemini 4 Argon
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
MiMo-V2.6-Pro vs Qwen4 Max: questions
- MiMo-V2.6-Pro vs Qwen4 Max: which is cheaper, and will that change?
- MiMo-V2.6-Pro is cheaper today: $0.544 against $2.25 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is MiMo-V2.6-Pro or Qwen4 Max better for coding?
- MiMo-V2.6-Pro scores higher in coding (58 vs 55 on the category index, where 50 is average). On FrontierCode 2.0 (AA), the top shared coding benchmark, it is 24.0% for MiMo-V2.6-Pro against 26.4% for Qwen4 Max.
- Which is faster, MiMo-V2.6-Pro or Qwen4 Max?
- Qwen4 Max streams faster (59 vs 56 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- MiMo-V2.6-Pro accepts more context: 1,048,576 against 991,808 tokens for Qwen4 Max.
- Which has lower latency?
- MiMo-V2.6-Pro answers first: 42.8 s against 139.2 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.
More questions (2)
- Do independent benchmarks agree on MiMo-V2.6-Pro vs Qwen4 Max?
- Across the 15 benchmarks both have, MiMo-V2.6-Pro leads 0, Qwen4 Max leads 15. Every overall index that rates both puts Qwen4 Max ahead.
- When was Qwen4 Max released?
- Qwen4 Max was released on 16 Nov 2026.