RankingEffort benchmarks

DeepSeek V3.1 Terminus

Compare models

28 measured results · 2 effort settings · 18 benchmark/harness combinations

Change model

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare with

Capability profile

One configuration, seven capabilities

Comparison score · 0–100
DeepSeek V3.1 Terminus · Reasoning capability radarComparison scores from 0 to 100. Unknown capabilities have no point; lines stop at gaps. Hollow points are preliminary. Exact values and support labels follow the chart.AgenticHard reasoningCodingHuman prefKnowledgeMultimodalLong context50100DeepSeek V3.1 Terminus · Reasoning · Agentic: 13.5 · SupportedDeepSeek V3.1 Terminus · Reasoning · Hard reasoning: 17.8 · SupportedDeepSeek V3.1 Terminus · Reasoning · Coding: 14.0 · SupportedDeepSeek V3.1 Terminus · Reasoning · Knowledge: 14.8 · SupportedDeepSeek V3.1 Terminus · Reasoning · Long context: 24.5 · Preliminary

Gaps are unknown, not zero. Hollow points are preliminary.

  • Agentic

    A13.5Supported
  • Hard reasoning

    A17.8Supported
  • Coding

    A14.0Supported
  • Human pref

    Unknown
  • Knowledge

    A14.8Supported
  • Multimodal

    Unknown
  • Long context

    A24.5Preliminary

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

DeepSeek V3.1 Terminus · Reasoning: 6.3 overall comparison score. Meets the overall evidence requirements.

5 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?
  • Artificial Analysis · 100.0% direct comparison weight · independent of the model provider
Which benchmark families carry the weight?
  • Gmmlu · 7.7% direct comparison weight
  • Terminal-Bench · 7.3% direct comparison weight
  • GDPval · 7.3% direct comparison weight
  • Tau Bench · 7.2% direct comparison weight
  • briefcase · 7.1% direct comparison weight
  • Other families (10) · 63.5% 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.

What changes when a source is removed?
Removed sourceRank before → afterRank movementCohort remaining / lost supportScore after (change)
Artificial AnalysisComparison support lost; no rank230 remaining / 372 lostUnavailable
ARC Prize302 → 300Up 2602 remaining / 0 lost7.9 (+1.6 points)
Datacurve302 → 302No movement602 remaining / 0 lost5.0 (-1.2 points)
LiveCodeBench302 → 302No movement602 remaining / 0 lost6.2 (-0.0 points)
Scale302 → 302No movement602 remaining / 0 lost7.0 (+0.7 points)
Terminal-Bench302 → 302No movement602 remaining / 0 lost6.2 (-0.0 points)
Cognition302 → 304Down 2602 remaining / 0 lost5.9 (-0.3 points)
LiveBench301 → 301No movement601 remaining / 1 lost6.6 (+0.3 points)
SWE-rebench302 → 302No movement602 remaining / 0 lost6.4 (+0.1 points)
Epoch AI295 → 296Down 1588 remaining / 14 lost6.5 (+0.2 points)
Vals AI302 → 302No movement601 remaining / 1 lost6.0 (-0.2 points)
Provider reports · Anthropic302 → 302No movement602 remaining / 0 lost6.3 (-0.0 points)
Provider reports · Moonshot AI302 → 302No movement602 remaining / 0 lost6.1 (-0.2 points)
Provider reports · Meta302 → 302No movement602 remaining / 0 lost6.2 (-0.0 points)
Provider reports · Mistral302 → 302No movement602 remaining / 0 lost6.3 (+0.0 points)
Provider reports · OpenAI302 → 302No movement602 remaining / 0 lost6.3 (+0.0 points)
All provider reports302 → 302No movement602 remaining / 0 lost6.1 (-0.1 points)

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 evidenceRecomputed scoreChange
arc-agi7.6+1.4
deepswe5.0-1.2
gmmlu6.7+0.5
mmmu-pro6.7+0.4
aa-lcr5.9-0.3
critpt5.9-0.3
frontiercode5.9-0.3
simpleqa6.5+0.3
epoch-game-puzzles6.0-0.3
aime6.0-0.2
terminal-bench6.5+0.2
hle6.5+0.2
automationbench6.4+0.2
livebench-instructions6.4+0.1
enigma-eval6.4+0.1
swe-rebench6.4+0.1
briefcase6.4+0.1
enterprise-ops6.1-0.1
multi-swe-bench6.1-0.1
arc-agi-36.2-0.1
apex-agents6.4+0.1
livebench-language6.4+0.1
ifbench6.2-0.1
tau-bench6.2-0.1
livecodebench6.2-0.1
mmlu-pro6.3+0.1
terminal-bench-science6.3+0.1
livebench-coding6.2-0.1
swe-atlas-qna6.2-0.1
gdpval6.3+0.1
vibe-code6.2-0.1
itbench6.2-0.1
math5006.3+0.0
vals-code-migration6.2-0.0
livebench-reasoning6.2-0.0
gpqa6.2-0.0
analyst-agent6.3+0.0
vals-legal-research6.2-0.0
gdp-pdf6.2-0.0
vals-finance-agent6.2-0.0
vals-tax-agent6.2-0.0
swe-atlas-test-writing6.2-0.0
vals-excel-modeling6.2-0.0
swe-bench-pro6.3+0.0
frontiermath6.2-0.0
omniscience6.3+0.0
livebench-data6.3+0.0
swe-atlas-refactoring6.3-0.0
harvey6.3+0.0
livebench-math6.3+0.0
What evidence is missing from the fit?

Across all collected settings for this model: 28 observations, 28 contributing, 0 matched without graph weight and 0 excluded. Counts do not establish rank eligibility.

    Inspect every observation and exclusion →

    Benchmark results

    These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

    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 / sourceReasoningNon-reasoning
    AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis

    Unit: elo. Source collected 2026-11-22.

    458.16 Elo
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    AIME 2025Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    89.67%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    53.67%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    69.33%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    45.33%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    2.71%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Benchmark 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.

    1.71%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    0%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    5.4%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis

    Unit: elo. Source collected 2026-11-22.

    717.53 Elo
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    Global-MMLU-LiteArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    82.69%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    79.19%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    75.05%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    16.36%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    8.67%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    57.01%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    41.22%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    LiveCodeBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    79.79%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    52.91%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    MMLU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    85.11%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    83.6%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    37.13%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    37.13%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    21.03%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    30.3%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    31.82%
    Reported settings

    DeepSeek V3.1 Terminus (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus

    Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    44.94%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    Not reported
    Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours
    Benchmark details

    Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    0%
    Reported settings

    DeepSeek V3.1 Terminus (Reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning

    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