RankingEffort benchmarks
Qwen3.6 Max Preview
11 measured results · 2 effort settings · 11 benchmark/harness combinations
Change model
Capability profile
One configuration, seven capabilities
- AQwen3.6 Max Preview · Thinking1 of 7 capabilities supported
Gaps are unknown, not zero. Hollow points are preliminary.
Agentic
UnknownHard reasoning
A32.5SupportedCoding
UnknownHuman pref
UnknownKnowledge
A59.3PreliminaryMultimodal
UnknownLong context
Unknown
Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.
Capability score evidence
Qwen3.6 Max Preview · Thinking: 29.7 overall comparison score. Meets the overall evidence requirements.
Which sources carry the weight?
- Epoch AI · 100.0% direct comparison weight · independent of the model provider
Which benchmark families carry the weight?
- GPQA · 29.1% direct comparison weight
- Aime · 28.7% direct comparison weight
- epoch-game-puzzles · 25.9% direct comparison weight
- simpleqa · 16.3% direct comparison weight
These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.
How sensitive is this score?
The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-11-22.
Both ranks use the same surviving configurations for each removal. The cohort starts with 602 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.
| Removed source | Rank before → after | Rank movement | Cohort remaining / lost support | Score after (change) |
|---|---|---|---|---|
| Artificial Analysis | 129 → 129 | No movement | 230 remaining / 372 lost | 28.5 |
| ARC Prize | 162 → 167 | Down 5 | 602 remaining / 0 lost | 32.7 |
| Datacurve | 162 → 159 | Up 3 | 602 remaining / 0 lost | 25.9 |
| LiveCodeBench | 162 → 162 | No movement | 602 remaining / 0 lost | 29.8 |
| Scale | 162 → 163 | Down 1 | 602 remaining / 0 lost | 31.3 |
| Terminal-Bench | 162 → 158 | Up 4 | 602 remaining / 0 lost | 30.2 |
| Cognition | 162 → 159 | Up 3 | 602 remaining / 0 lost | 29.7 |
| LiveBench | 161 → 160 | Up 1 | 601 remaining / 1 lost | 29.9 |
| SWE-rebench | 162 → 160 | Up 2 | 602 remaining / 0 lost | 29.7 |
| Epoch AI | Comparison support lost; no rank | 588 remaining / 14 lost | Unavailable | |
| Vals AI | 162 → 163 | Down 1 | 601 remaining / 1 lost | 28.4 |
| Provider reports · Anthropic | 162 → 162 | No movement | 602 remaining / 0 lost | 29.7 |
| Provider reports · Moonshot AI | 162 → 162 | No movement | 602 remaining / 0 lost | 29.3 |
| Provider reports · Meta | 162 → 162 | No movement | 602 remaining / 0 lost | 29.8 |
| Provider reports · Mistral | 162 → 162 | No movement | 602 remaining / 0 lost | 29.7 |
| Provider reports · OpenAI | 162 → 162 | No movement | 602 remaining / 0 lost | 29.8 |
| All provider reports | 162 → 162 | No movement | 602 remaining / 0 lost | 29.5 |
Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.
Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit
Every benchmark family
| Removed evidence | Recomputed score | Change |
|---|---|---|
| simpleqa | 24.3 | -5.5 |
| epoch-game-puzzles | 34.3 | +4.6 |
| deepswe | 26.0 | -3.8 |
| arc-agi | 31.8 | +2.1 |
| gpqa | 31.6 | +1.9 |
| mmmu-pro | 30.8 | +1.1 |
| terminal-bench | 30.3 | +0.6 |
| hle | 30.2 | +0.5 |
| mmlu-pro | 30.2 | +0.5 |
| enigma-eval | 30.2 | +0.5 |
| tau-bench | 29.3 | -0.4 |
| aime | 30.2 | +0.4 |
| livebench-instructions | 30.1 | +0.4 |
| frontiermath | 30.0 | +0.3 |
| multi-swe-bench | 29.4 | -0.3 |
| swe-atlas-qna | 29.4 | -0.3 |
| livebench-coding | 29.4 | -0.3 |
| livebench-language | 30.0 | +0.3 |
| ifbench | 30.0 | +0.2 |
| vibe-code | 29.6 | -0.1 |
| gdp-pdf | 29.9 | +0.1 |
| livebench-reasoning | 29.6 | -0.1 |
| vals-legal-research | 29.6 | -0.1 |
| apex-agents | 29.9 | +0.1 |
| vals-finance-agent | 29.6 | -0.1 |
| automationbench | 29.6 | -0.1 |
| terminal-bench-science | 29.8 | +0.1 |
| gdpval | 29.6 | -0.1 |
| briefcase | 29.6 | -0.1 |
| livecodebench | 29.8 | +0.1 |
| enterprise-ops | 29.6 | -0.1 |
| itbench | 29.6 | -0.1 |
| swe-atlas-test-writing | 29.6 | -0.1 |
| vals-tax-agent | 29.6 | -0.1 |
| vals-code-migration | 29.6 | -0.1 |
| livebench-data | 29.8 | +0.1 |
| critpt | 29.7 | -0.1 |
| arc-agi-3 | 29.8 | +0.1 |
| analyst-agent | 29.8 | +0.1 |
| vals-excel-modeling | 29.7 | -0.0 |
| gmmlu | 29.8 | +0.0 |
| livebench-math | 29.8 | +0.0 |
| swe-bench-pro | 29.7 | +0.0 |
| omniscience | 29.7 | -0.0 |
| swe-atlas-refactoring | 29.7 | -0.0 |
| swe-rebench | 29.7 | -0.0 |
| frontiercode | 29.7 | +0.0 |
| math500 | 29.7 | +0.0 |
| harvey | 29.7 | +0.0 |
| aa-lcr | 29.7 | -0.0 |
What evidence is missing from the fit?
Across all collected settings for this model: 11 observations, 11 contributing, 0 matched without graph weight and 0 excluded. Counts do not establish rank eligibility.
Benchmark results
Reviewed retained observations are in use. October 2026 Real-SWE board has no 95% confidence whisker, so these scores stay at the 2026-10-18 retrieval. Original source retrieval 2026-10-18; source updated 2026-10-18.
| Benchmark / source | Thinking | Reasoning |
|---|---|---|
AA-LCR v1.1Benchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 80.67% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
CritPtBenchmark detailsBenchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 3.71% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
GPQA DiamondBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 88.79% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
Humanity's Last ExamBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 30.82% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
IFBenchBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 76.6% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
𝜏²-Bench TelecomBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 95.91% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
Terminal-Bench HardBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | Not reported | 43.94% Reported settingsQwen3.6 Max Preview; Reported effort: reasoning; Model source: https://artificialanalysis.ai/models/qwen3-6-max; Mode from evaluator isReasoning metadata; effort budget not specified |
Epoch · GPQA diamond · v1.0.11Benchmark detailsEpoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-11-22. | 87.37% Reported settingsEpoch AI Inspect; task 1.0.11; scorer choice; model=qwen3.6-max-preview; run=Jv2hJvjpvA8xJW3Bzv7bfw; mean accuracy; https://epoch.ai/data/benchmarks.csv | Not reported |
Epoch · Mystery Game Puzzles · v1.0.4 · Tool submissionBenchmark detailsEpoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-11-22. | 19% Reported settingsEpoch AI Inspect; task 1.0.4; scorer agent_submitted_next_move_scorer; model=qwen3.6-max-preview_none; run=mbySBWyYrKdtd7joxHyBdq; mean accuracy; https://epoch.ai/data/benchmarks.csv | Not reported |
Epoch · OTIS Mock AIME 2024-2025 · v1.1.6Benchmark detailsEpoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-11-22. | 91.11% Reported settingsEpoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=qwen3.6-max-preview; run=EfJpx5HztMmL5EQSwsyqMe; mean accuracy; https://epoch.ai/data/benchmarks.csv | Not reported |
Epoch · SimpleQA Verified · v1.2.0Benchmark detailsEpoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-11-22. | 52% Reported settingsEpoch AI Inspect; task 1.2.0; scorer simpleqa_scorer; model=qwen3.6-max-preview; run=iyPiR8GUuyNNA6Fssziq32; mean accuracy; https://epoch.ai/data/benchmarks.csv | Not reported |
What these comparisons mean
Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.
These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.
Download all effort benchmark results · 10,184 results across 381 models · Collected 2026-11-22