Skip to main content
BenchLM
Data

Data as of November 17, 2026 · How the score is built

Superseded.Alibaba has newer models in this line:Qwen3.8 Max Preview

Qwen3.7 Max

SupersededProprietaryReasoning

Decision readingQwen3.7 Max scores 60.8 out of 100 and ranks #50 of 216. This profile shows 48 source-displayable benchmark rows; its strongest eligible category is Multilingual at #1. No comparable first-party API price is published in the catalog.

Email me when its price, rank or availability changes.

Capability

60.8/100

field median 49.4#50 of 216 ranked models

Public

#50of 216

Verified #32 of 80

Price

API rate not published

input median $0.60No comparable first-party hosted token rate

Speed

197tok/s

field median 89.5 tok/sFirst token 14.47 s

Context

1Mtokens

field median 256,000Reported for this model; direct source link not stored

Decision snapshot

Where this model sits in the ranked field, its strongest published evidence, and what to validate before choosing it. The marker compares it with the current ranked catalog, not with an absolute quality threshold.

Strongest published evidence

Multilingual ranks #1. A well-rounded choice across a range of tasks.

Validate before choosing

48 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.

Source-linked · 48 displayable benchmark rows

Is Qwen3.7 Max right for your job?

Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

Category score record

Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #75 of 138Percentile 46thWeight 22%10 benchmarksVerified
40.1
CodingRank #62 of 163Percentile 62ndWeight 20%10 benchmarksVerified
44.8
ReasoningRank #48 of 160Percentile 70thWeight 17%2 benchmarksVerified
51.7
MultimodalWeight 12%0 benchmarksNot measured
Not measured
KnowledgeRank #48 of 181Percentile 74thWeight 12%9 benchmarksVerified
60.2
MultilingualRank #1 of 17Percentile 100thWeight 7%5 benchmarksVerified
100.0
Inst. FollowingRank #17 of 157Percentile 90thWeight 5%2 benchmarksVerified
92.8
MathRank #2 of 12Percentile 91stWeight 5%5 benchmarksVerified
78.3

48 of 922 tracked benchmark slots have displayable evidence · bars run 0–100

Coverage details

How much of this is verified

Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.

  1. Agentic10/10 verified
  2. Coding10/10 verified
  3. Reasoning2/2 verified
  4. MultimodalNot measured
  5. Knowledge9/9 verified
  6. Multilingual5/5 verified
  7. Inst. Following2/2 verified
  8. Math5/5 verified
Verified sourceProvisionalNot measured

Capability shape

Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.

Qwen3.7 Max category percentile values

  • Agentic46th percentile
  • Coding62nd percentile
  • Reasoning70th percentile
  • MultimodalNot eligible
  • Knowledge74th percentile
  • Multilingual100th percentile
  • Instruction following90th percentile
  • Math91st percentile

The dashed outline is median of 6 nearest peers.

Eligible category ranks

  1. Agentic#75/138
  2. Coding#62/163
  3. Reasoning#48/160
  4. MultimodalNot ranked
  5. Knowledge#48/181
  6. Multilingual#1/17
  7. Inst. Following#17/157
  8. Math#2/12
Top decileTop quartileMid-fieldNot eligible

Benchmark ledger

Every result says who ran the test. For this model, Alibaba ran 33 of 48. Coding opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.

Overall rank
#50
of 216 on the leaderboard
Reported by Alibaba
33
Shown with that label
Run by a benchmark
10
The benchmark’s own run
Independent runs
5
An outside evaluator’s run
Coding10 rows
Coding benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
SWE-bench Pro60.6%

Best verified: Claude Fable 5.2 · 93%

Gap32.4 behindReported by AlibabaCounts26% ref. weight
SciCodeScientific Code Benchmark53.5%

Best verified: Sakana Fugu · 60.1%

Gap6.6 behindReported by AlibabaCounts10% ref. weight
LiveCodeBench (Vals)LiveCodeBench, Vals AI run87.1%

Best verified: Claude Fable 5.1 · 90.5%

Gap3.4 behindIndependent run: Vals AICounts8% ref. weight
SWE Multilingual78.3%

Best verified: Claude Fable 5.2 · 96.6%

Gap18.3 behindReported by AlibabaCounts5% ref. weight
SWE-bench VerifiedSoftware Engineering Benchmark Verified80.4%

Best verified: Claude Opus 5 · 96%

Gap15.6 behindReported by AlibabaShown only
NL2Repo47.2%

Best verified: DeepSeek V4.1 Flash 1017 · 68.2%

Gap21 behindReported by AlibabaShown only
LiveCodeBench v691.6%

Best verified: Sakana Fugu-Ultra · 93.2%

Gap1.6 behindReported by AlibabaShown only
Terminal-Bench 2.069.7%

Best verified: GPT-5.5 · 82.0%

Gap12.3 behindReported by AlibabaShown only
OpenHarmony Bench v1.053.4%

Best verified: Qwen3.8 Max · 60.8%

Gap7.4 behindRun by OpenHarmony Bench v1.0Shown only
SWE-bench (Vals)SWE-bench, Vals AI run68.8%

Best verified: Claude Opus 5 · 97.0%

Gap28.2 behindIndependent run: Vals AIShown only
Agentic10 rows
Agentic benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
MCP Atlas76.4%

Best verified: Muse Spark Open · 89.2%

Gap12.8 behindReported by AlibabaCounts5% ref. weight
Terminal-Bench 2.1 (Vals)Terminal-Bench 2.1, Vals AI run61.0%

Best verified: Claude Opus 5.5 · 87.6%

Gap26.6 behindIndependent run: Vals AICounts4% ref. weight
BFCL v4Berkeley Function Calling Leaderboard v475.0%

Best verified: BTL-3 · 88.5%

Gap13.5 behindReported by AlibabaCounts3% ref. weight
HLE w/ toolsHumanity's Last Exam with tools53.5%

Best verified: Kimi K3.1 Pro · 70.0%

Gap16.5 behindReported by AlibabaCounts3% ref. weight
Terminal-Bench 2.069.7%

Best verified: GPT-5.5 · 82%

Gap12.3 behindReported by AlibabaShown only
QwenClawBench64.3%

Best verified: Qwen3.7 Max · 64.3%

GapHighest listedReported by AlibabaShown only
Claw-Eval65.2%

Best verified: Ornith-1.5-397B · 81.4%

Gap16.2 behindReported by AlibabaShown only
VITA-Bench47.9%

Best verified: Qwen3.7 Max · 47.9%

GapHighest listedReported by AlibabaShown only
Gert LabsGert Labs Composite Game Benchmark64.27%

Best verified: Claude Opus 4.8 · 72.97%

Gap8.7 behindRun by Gert LabsShown only
ResearchClawBench18.7%

Best verified: Claude Opus 4.8 · 21.1%

Gap2.4 behindRun by ResearchClawBenchShown only
Reasoning2 rows
Reasoning benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
CritPtCritical Physics Tasks13.4%

Best verified: Claude Fable 5.2 · 35.4%

Gap22 behindReported by AlibabaCounts25% ref. weight
MRCRv290.4%

Best verified: Sakana Fugu-Ultra · 93.6%

Gap3.2 behindReported by AlibabaCounts20% ref. weight
Knowledge9 rows
Knowledge benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
HLEHumanity's Last Exam41.4%

Best verified: Claude Fable 5.2 · 74.2%

Gap32.8 behindReported by AlibabaCounts44% ref. weight
MMLU-ProMassive Multitask Language Understanding Professional89.6%

Best verified: Qwen3.7 Max · 89.6%

GapHighest listedReported by AlibabaCounts6% ref. weight
MMLU-Pro (Vals)MMLU-Pro, Vals AI run89.3%

Best verified: Claude Fable 5.1 · 92.4%

Gap3.1 behindIndependent run: Vals AICounts6% ref. weight
GPQAGraduate-Level Google-Proof Q&A92.4%

Best verified: GPT-6 Astra · 96%

Gap3.6 behindReported by AlibabaCounts3% ref. weight
SuperGPQASuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines73.6%

Best verified: Qwen 3.6 Max (preview) · 73.9%

Gap0.3 behindReported by AlibabaCounts2% ref. weight
GPQA Diamond (Vals)GPQA Diamond, Vals AI run90.2%

Best verified: Gemini 3.1 Pro · 95.5%

Gap5.3 behindIndependent run: Vals AICounts2% ref. weight
GPQA-DGPQA Diamond92.4%

Best verified: GPT-6 Astra · 96.0%

Gap3.6 behindReported by AlibabaShown only
MMLU-Redux95%

Best verified: Qwen3.7 Max · 95%

GapHighest listedReported by AlibabaShown only
MMMLU90.3%

Best verified: Interfaze Beta · 90.9%

Gap0.6 behindReported by AlibabaShown only
Multilingual5 rows
Multilingual benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
MMLU-ProX87%

Best verified: Qwen3.7 Max · 87%

GapHighest listedReported by AlibabaCountsWeighted 100%
NOVA-6359.0%

Best verified: Qwen3.5 397B · 59.1%

Gap0.1 behindReported by AlibabaShown only
INCLUDE86.2%

Best verified: Claude Opus 5 · 89.8%

Gap3.6 behindReported by AlibabaShown only
MAXIFE89.2%

Best verified: Qwen3.7 Max · 89.2%

GapHighest listedReported by AlibabaShown only
PolyMath86.5%

Best verified: Qwen3.7 Max · 86.5%

GapHighest listedReported by AlibabaShown only
Inst. Following2 rows
Inst. Following benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
IFBenchInstruction Following Benchmark79.1%

Best verified: Flint · 89.2%

Gap10.1 behindReported by AlibabaCountsWeighted 70%
IFEvalInstruction-Following Eval94.3%

Best verified: LFM2.5-24B-A2B · 96.8%

Gap2.5 behindReported by AlibabaShown only
Math5 rows
Math benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
FrontierMath v2 (Tiers 1-3)FrontierMath v2 Tiers 1-364.561%

Best verified: Claude Fable 5.2 · 95.088%

Gap30.5 behindRun by FrontierMath v2 (Tiers 1-3)CountsWeighted 30%
HMMT Feb 2026Harvard-MIT Mathematics Tournament February 202697.1%

Best verified: Qwen3.7 Max · 97.1%

GapHighest listedReported by AlibabaCountsWeighted 25%
FrontierMath v2 (Tier 4)FrontierMath v2 Tier 434.146%

Best verified: Claude Fable 5.2 · 100.000%

Gap65.9 behindRun by FrontierMath v2 (Tier 4)CountsWeighted 10%
IMOAnswerBench90.0%

Best verified: dots3-note Preview · 90.9%

Gap0.9 behindReported by AlibabaShown only
Apex44.5%

Best verified: Hy4 preview · 74.2%

Gap29.7 behindReported by AlibabaShown only
External signals5 rows
External signals benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
NYT ConnectionsNYT Connections LLM Benchmark (extended version)85.1%

Best verified: Gemini 4 Carbon · 99.3%

Gap14.2 behindRun by NYT ConnectionsShown only
BabaIsBenchBabaIsBench (Baba Is You intro levels, pass@1)25%

Best verified: Claude Opus 5 · 100%

Gap75 behindRun by BabaIsBenchShown only
Chess PuzzlesChess Puzzles (Epoch AI)27.0%

Best verified: Claude Fable 5.2 · 79.0%

Gap52 behindRun by Chess PuzzlesShown only
Mystery Game PuzzlesMystery Game Puzzles (Epoch AI)32.0%

Best verified: Claude Fable 5.2 · 90.0%

Gap58 behindRun by Mystery Game PuzzlesShown only
EBR-benchEarthborne Rangers (EBR-bench)9.5%

Best verified: Claude Fable 5.2 · 90.5%

Gap81 behindRun by EBR-benchShown only

Outlined bars are lab-reported. Filled bars are benchmark-run or independent. Bars run 0–100; the dark tick marks the best source-verified value.

Lineage

The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.

  1. May 16, 2026 · you are here

    Qwen3.7 Max

    Score 60.8 · Price not listed

  2. Jul 19, 2026

    Qwen3.8 Max Preview

    Not publicly ranked · Price not listed

Base entry

Spec sheet

Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.

API model ID
Not published
Context window
1M
Maximum output
Not sourced yet
Knowledge cutoff
Not sourced yet
Parameters
Not disclosed by the provider
Availability
Not sourced yet
Cloud regions
Not tracked yet
Lifecycle
Superseded
API capabilities
Tool calling, structured outputs, and batch support are not tracked yet
Prompt caching
Not documented in the pricing record
Self-host
Weights are not published
Rate limits
Not tracked yet

How to read this profile

The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.

Qwen3.7 Max ranks #50 of 216 on the public leaderboard with a score of 60.78/100. Its source-verified position is #32 of 80.

Qwen3.7 Max is a proprietary model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.

Official exact-value snapshot from Qwen's May 16, 2026 Qwen3.7 launch post. The post names the API model as qwen3.7-max, describes a 1,000,000-token context window with thinking enabled, and says Model Studio availability is coming soon.

48 of 922 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.

Its strongest eligible category is Multilingual at #1, while its lowest eligible position is Agentic at #75. a well-rounded choice across a range of tasks.

Last updated November 17, 2026. Runtime fields remain blank until a sourced snapshot exists.

Deployment options

Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.

Where to run Qwen3.7 Max by API

4 hosts publish a per-token price for it. Novita AI has the lowest output rate, $3.75 per million tokens. Prices read from each host’s own pages on November 16, 2026.

Inference hosts for Qwen3.7 Max with input, output and cached-input prices per million tokens.
HostInput /1MOutput /1MCached /1MContext
Novita AI$1.25$3.75$0.251M
OpenRouter$1.48$4.42$0.291M
DeepInfra$2.50$7.50$0.50256K
Together AI$2.50$7.50$0.50—
Compare inference hosts

Questions

How does Qwen3.7 Max perform overall in AI benchmarks?

Qwen3.7 Max ranks #50 out of 216 models on the public BenchAlign leaderboard, with a score of 60.78/100. Its evidence status is Supported, and this profile shows 48 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.

Is Qwen3.7 Max good for knowledge and understanding?

Qwen3.7 Max ranks #48 out of 181 eligible models for knowledge and understanding, with a public category score of 60.2/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for coding and programming?

Qwen3.7 Max ranks #62 out of 163 eligible models for coding and programming, with a public category score of 44.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for mathematics?

Qwen3.7 Max ranks #2 out of 12 eligible models for mathematics, with a public category score of 78.3/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for reasoning and logic?

Qwen3.7 Max ranks #48 out of 160 eligible models for reasoning and logic, with a public category score of 51.7/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for agentic tool use and computer tasks?

Qwen3.7 Max ranks #75 out of 138 eligible models for agentic tool use and computer tasks, with a public category score of 40.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for instruction following?

Qwen3.7 Max ranks #17 out of 157 eligible models for instruction following, with a public category score of 92.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Qwen3.7 Max good for multilingual tasks?

Qwen3.7 Max ranks #1 out of 17 eligible models for multilingual tasks, with a public category score of 100/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.

Does Qwen3.7 Max have full benchmark coverage on BenchLM?

No. Qwen3.7 Max currently has 67 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.

What is the context window size of Qwen3.7 Max?

Qwen3.7 Max has a reported context window of 1M in the exact-model catalog record. The value stays visible, but the profile marks its source link as unavailable instead of presenting it as directly documented. Maximum output length remains separate because providers often publish a different limit.

Compare Qwen3.7 Max with every tracked model1039 comparisons