UnifyBench

RankingQwen3-235B-A22B-Instruct-2507

Qwen3-235B-A22B-Instruct-2507

Data updated 12 Sept 2026

3 published benchmark measures · 3 benchmark families contribute across 3 task areas. 2 capability estimates available in the full profile. See all results ↓ · Compare published benchmarks →

Model evidence summary

Choose a configuration

A model-wide score would mix effort settings. This page summarizes published evidence; only identified configurations receive leaderboard ranks.

Compare measured configurations →

The capability chart summarizes evidence across reported settings, not a runnable configuration or a leaderboard rank.

Performance profile

Capabilities

Adjusted comparison score · 0–100
Qwen3-235B-A22B-Instruct-2507 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 context50100Qwen3-235B-A22B-Instruct-2507 · Hard reasoning: 5.8 · PreliminaryQwen3-235B-A22B-Instruct-2507 · Human pref: 14.0 · Supported

Filled points are supported; hollow points are preliminary. Lines stop at unknown capabilities. Exact values and sources follow below.

AgenticNo comparable evidenceUnknown

0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 11/11 scenarios unsupported. Without one publisher: No supported estimate; 15/15 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage; source sensitivity still requires inspection. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

    Hard reasoning1 families · 1 with independent evidence · Preliminary5.8

    1 core families; 47 direct opponents across 9 labs. Needs broader benchmark and opponent support

    Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

    Refit without one family: No supported estimate; 7/7 scenarios unsupported. Without one publisher: No supported estimate; 16/16 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

    Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

    Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

    • ARC-AGI: 13.8 observed win share
      arcprize.org · Source 1
    CodingNo comparable evidenceUnknown

    0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

    Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

    Refit without one family: No supported estimate; 7/7 scenarios unsupported. Without one publisher: No supported estimate; 20/20 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

    Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

    Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

      Human pref1 families · 1 with independent evidence · Supported14.0

      1 core families; 114 direct opponents across 19 labs.

      Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

      Refit without one family: No supported estimate; 1/1 scenarios unsupported. Without one publisher: No supported estimate; 1/1 unsupported. Smoothing check: 5.4–27.3; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

      Audit: cross-family and cross-publisher predictions could not be evaluated from the available comparison network. Validation audit.

      Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

      • Human preference (Arena): 62.3 observed win share
        lmarena.ai · Source 1
      Knowledge1 families · 1 with independent evidence · PreliminaryUnknown

      1 core families; 51 direct opponents across 12 labs. No connected comparison to the complete reference panel

      Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

      Refit without one family: No supported estimate; 4/4 scenarios unsupported. Without one publisher: No supported estimate; 5/5 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

      Audit: almost no publisher-held-out comparisons remained estimable. Cross-source validity is unresolved. Validation audit.

      Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

      • MMLU-Pro: 72.5 observed win share
        huggingface.co · Source 1
      MultimodalNo comparable evidenceUnknown

      0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

      Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

      Refit without one family: No supported estimate; 8/8 scenarios unsupported. Without one publisher: No supported estimate; 8/8 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

      Audit: limited publisher-held-out coverage. Cross-source validity is unresolved. Validation audit.

      Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

        Long contextNo comparable evidenceUnknown

        0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

        Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

        Refit without one family: No supported estimate; 5/5 scenarios unsupported. Without one publisher: No supported estimate; 5/5 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

        Audit: family and publisher holdouts underperformed neutral predictions. Treat this as an exploratory estimate. Validation audit.

        Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

          Scores estimate outcomes against a shared reference panel; they are not accuracy percentages. Sparse or disconnected evidence cannot qualify an overall profile. Open a capability to inspect its evidence.

          Compare 1 effort levels across 13 benchmark/harness combinations →

          Reported effort · Not specified

          Settings reported in this model's published benchmark results, including results outside the aggregate. Effort names are provider-specific. These are not API defaults or equal compute budgets.

          Mixed settings means multiple settings occur in the evidence. Best across efforts means the source selected its best reported result across settings; it does not mean Max. Unspecified settings stay unknown. This model-summary chart combines reported settings. The leaderboard keeps identified configurations separate and excludes unknown effort.

          Inspect each result and its source ↓ · Download effort evidence

          Compare capability profiles →

          Score contributions and missing evidence

          3 contributing families across 3 capabilities. Fixed reference panels do not change when the catalog expands.

          Capability is fitted jointly across families. Capability estimates below describe different task areas; their weighted sum is not the Capability score.

          Results without reviewed compatibility or a reference match remain in the raw evidence below. Coverage counts only contributing results.

          Model information & shareable badge
          Lab
          Qwen
          Catalog status
          open-weights-available
          Availability
          Open weights available for self-hosting
          Family
          Qwen3
          Released
          —
          Context
          —
          License
          apache-2.0
          Model card
          https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507
          Default Capability family coverage
          Documented-evidence family coverage/badge/qwen3-235b-a22b-instruct-2507.svg
          Benchmark scores & sources

          Original results, evaluation harnesses, and evidence behind this model.

          BenchmarkBucketScoreHarnessEvidenceSource-recorded date
          ARC-AGI-2Hard reasoning1.3%ARC Prize verified

          Effort: Not specified

          contributes to capability          		official board2025-07-25
          LMArena Text ArenaHuman pref1423LMArena Text

          Effort: Not specified

          contributes to capability          		official board2026-09-11
          MMLU-ProKnowledge83%MMLU-Pro reported

          Effort: Not specified

          contributes to capability          		official board2026-09-11
          DeepSWE v1.1Agentic————
          GDPval-AAAgentic————
          GPQA DiamondHard reasoning————
          Humanity's Last ExamHard reasoning————
          LiveCodeBenchCoding————
          OSWorld-VerifiedAgentic————
          SWE-bench ProAgentic————
          SWE-bench VerifiedAgentic————
          Terminal-Bench 2.1Agentic————