Microsoft: Phi 4

[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 …

microsoft/phi-4

[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion...

Specifications & capabilities

Input
text
Output
text
Context window · tokens
16384
Knowledge cutoff
2024-06-30
Published model ID
phi-4
Capabilities
structuredOutputs

Evaluations

1–20 / 83

Open LLM Leaderboard v2

40.604%

source id
openLLMLeaderboardV2
leaderboard name
Open LLM v2 GPQA · Raw
dimension
knowledge
metric
Raw accuracy
rank
25
aggregation scope
suite
Open LLM v2
co2_kg
2.7731595658140327
status
archived
precision
bfloat16
task_count
6
metric_kind
Raw accuracy
rank_method
Derived competition rank over published evaluations with reported scores, including variants and source flags
weight_type
Original
published_at
2025-03-20T12:17:27Z
chat_template
true
suite_version
2
model_revision
381727a5ee103da6c1b14ecd3d39cd09832cbcf8
source_flagged
false
missing_metrics
    dataset_revision
    9c09a7cae43334062a82cb164f2ef255013dafa2
    metric_population
    4576
    ranking_population
    4576
    source_flag_anomaly
    false
    source model id
    microsoft/phi-4
    Original result ↗

    Open LLM Leaderboard v2

    20.805%

    source id
    openLLMLeaderboardV2
    leaderboard name
    Open LLM v2 GPQA · Normalized
    dimension
    knowledge
    metric
    Normalized score
    rank
    25
    aggregation scope
    suite
    Open LLM v2
    co2_kg
    2.7731595658140327
    status
    archived
    precision
    bfloat16
    task_count
    6
    metric_kind
    Normalized score
    rank_method
    Derived competition rank over published evaluations with reported scores, including variants and source flags
    weight_type
    Original
    published_at
    2025-03-20T12:17:27Z
    chat_template
    true
    suite_version
    2
    model_revision
    381727a5ee103da6c1b14ecd3d39cd09832cbcf8
    source_flagged
    false
    missing_metrics
      dataset_revision
      9c09a7cae43334062a82cb164f2ef255013dafa2
      metric_population
      4576
      ranking_population
      4576
      source_flag_anomaly
      false
      source model id
      microsoft/phi-4
      Original result ↗

      Open LLM Leaderboard v2

      40.101%

      source id
      openLLMLeaderboardV2
      leaderboard name
      Open LLM v2 GPQA · Raw
      dimension
      knowledge
      metric
      Raw accuracy
      rank
      38
      aggregation scope
      suite
      Open LLM v2
      co2_kg
      0.8783618221059865
      status
      archived
      precision
      float16
      task_count
      6
      metric_kind
      Raw accuracy
      rank_method
      Derived competition rank over published evaluations with reported scores, including variants and source flags
      weight_type
      Original
      published_at
      2025-03-20T12:17:27Z
      chat_template
      false
      suite_version
      2
      model_revision
      381727a5ee103da6c1b14ecd3d39cd09832cbcf8
      source_flagged
      false
      missing_metrics
        dataset_revision
        9c09a7cae43334062a82cb164f2ef255013dafa2
        metric_population
        4576
        ranking_population
        4576
        source_flag_anomaly
        false
        source model id
        microsoft/phi-4
        Original result ↗

        Open LLM Leaderboard v2

        20.134%

        source id
        openLLMLeaderboardV2
        leaderboard name
        Open LLM v2 GPQA · Normalized
        dimension
        knowledge
        metric
        Normalized score
        rank
        38
        aggregation scope
        suite
        Open LLM v2
        co2_kg
        0.8783618221059865
        status
        archived
        precision
        float16
        task_count
        6
        metric_kind
        Normalized score
        rank_method
        Derived competition rank over published evaluations with reported scores, including variants and source flags
        weight_type
        Original
        published_at
        2025-03-20T12:17:27Z
        chat_template
        false
        suite_version
        2
        model_revision
        381727a5ee103da6c1b14ecd3d39cd09832cbcf8
        source_flagged
        false
        missing_metrics
          dataset_revision
          9c09a7cae43334062a82cb164f2ef255013dafa2
          metric_population
          4576
          ranking_population
          4576
          source_flag_anomaly
          false
          source model id
          microsoft/phi-4
          Original result ↗

          Open LLM Leaderboard v2

          50.335%

          source id
          openLLMLeaderboardV2
          leaderboard name
          Open LLM v2 MUSR · Raw
          dimension
          reasoning
          metric
          Raw accuracy
          rank
          60
          aggregation scope
          suite
          Open LLM v2
          co2_kg
          2.7731595658140327
          status
          archived
          precision
          bfloat16
          task_count
          6
          metric_kind
          Raw accuracy
          rank_method
          Derived competition rank over published evaluations with reported scores, including variants and source flags
          weight_type
          Original
          published_at
          2025-03-20T12:17:27Z
          chat_template
          true
          suite_version
          2
          model_revision
          381727a5ee103da6c1b14ecd3d39cd09832cbcf8
          source_flagged
          false
          missing_metrics
            dataset_revision
            9c09a7cae43334062a82cb164f2ef255013dafa2
            metric_population
            4576
            ranking_population
            4576
            source_flag_anomaly
            false
            source model id
            microsoft/phi-4
            Original result ↗

            Open LLM Leaderboard v2

            50.335%

            source id
            openLLMLeaderboardV2
            leaderboard name
            Open LLM v2 MUSR · Raw
            dimension
            reasoning
            metric
            Raw accuracy
            rank
            60
            aggregation scope
            suite
            Open LLM v2
            co2_kg
            0.8783618221059865
            status
            archived
            precision
            float16
            task_count
            6
            metric_kind
            Raw accuracy
            rank_method
            Derived competition rank over published evaluations with reported scores, including variants and source flags
            weight_type
            Original
            published_at
            2025-03-20T12:17:27Z
            chat_template
            false
            suite_version
            2
            model_revision
            381727a5ee103da6c1b14ecd3d39cd09832cbcf8
            source_flagged
            false
            missing_metrics
              dataset_revision
              9c09a7cae43334062a82cb164f2ef255013dafa2
              metric_population
              4576
              ranking_population
              4576
              source_flag_anomaly
              false
              source model id
              microsoft/phi-4
              Original result ↗

              Open LLM Leaderboard v2

              23.786%

              source id
              openLLMLeaderboardV2
              leaderboard name
              Open LLM v2 MUSR · Normalized
              dimension
              reasoning
              metric
              Normalized score
              rank
              65
              aggregation scope
              suite
              Open LLM v2
              co2_kg
              2.7731595658140327
              status
              archived
              precision
              bfloat16
              task_count
              6
              metric_kind
              Normalized score
              rank_method
              Derived competition rank over published evaluations with reported scores, including variants and source flags
              weight_type
              Original
              published_at
              2025-03-20T12:17:27Z
              chat_template
              true
              suite_version
              2
              model_revision
              381727a5ee103da6c1b14ecd3d39cd09832cbcf8
              source_flagged
              false
              missing_metrics
                dataset_revision
                9c09a7cae43334062a82cb164f2ef255013dafa2
                metric_population
                4576
                ranking_population
                4576
                source_flag_anomaly
                false
                source model id
                microsoft/phi-4
                Original result ↗

                Open LLM Leaderboard v2

                23.719%

                source id
                openLLMLeaderboardV2
                leaderboard name
                Open LLM v2 MUSR · Normalized
                dimension
                reasoning
                metric
                Normalized score
                rank
                70
                aggregation scope
                suite
                Open LLM v2
                co2_kg
                0.8783618221059865
                status
                archived
                precision
                float16
                task_count
                6
                metric_kind
                Normalized score
                rank_method
                Derived competition rank over published evaluations with reported scores, including variants and source flags
                weight_type
                Original
                published_at
                2025-03-20T12:17:27Z
                chat_template
                false
                suite_version
                2
                model_revision
                381727a5ee103da6c1b14ecd3d39cd09832cbcf8
                source_flagged
                false
                missing_metrics
                  dataset_revision
                  9c09a7cae43334062a82cb164f2ef255013dafa2
                  metric_population
                  4576
                  ranking_population
                  4576
                  source_flag_anomaly
                  false
                  source model id
                  microsoft/phi-4
                  Original result ↗

                  Open LLM Leaderboard v2

                  52.576%

                  source id
                  openLLMLeaderboardV2
                  leaderboard name
                  Open LLM v2 BBH · Normalized
                  dimension
                  reasoning
                  metric
                  Normalized score
                  rank
                  209
                  aggregation scope
                  suite
                  Open LLM v2
                  co2_kg
                  0.8783618221059865
                  status
                  archived
                  precision
                  float16
                  task_count
                  6
                  metric_kind
                  Normalized score
                  rank_method
                  Derived competition rank over published evaluations with reported scores, including variants and source flags
                  weight_type
                  Original
                  published_at
                  2025-03-20T12:17:27Z
                  chat_template
                  false
                  suite_version
                  2
                  model_revision
                  381727a5ee103da6c1b14ecd3d39cd09832cbcf8
                  source_flagged
                  false
                  missing_metrics
                    dataset_revision
                    9c09a7cae43334062a82cb164f2ef255013dafa2
                    metric_population
                    4576
                    ranking_population
                    4576
                    source_flag_anomaly
                    false
                    source model id
                    microsoft/phi-4
                    Original result ↗

                    Open LLM Leaderboard v2

                    67.035%

                    source id
                    openLLMLeaderboardV2
                    leaderboard name
                    Open LLM v2 BBH · Raw
                    dimension
                    reasoning
                    metric
                    Raw accuracy
                    rank
                    210
                    aggregation scope
                    suite
                    Open LLM v2
                    co2_kg
                    0.8783618221059865
                    status
                    archived
                    precision
                    float16
                    task_count
                    6
                    metric_kind
                    Raw accuracy
                    rank_method
                    Derived competition rank over published evaluations with reported scores, including variants and source flags
                    weight_type
                    Original
                    published_at
                    2025-03-20T12:17:27Z
                    chat_template
                    false
                    suite_version
                    2
                    model_revision
                    381727a5ee103da6c1b14ecd3d39cd09832cbcf8
                    source_flagged
                    false
                    missing_metrics
                      dataset_revision
                      9c09a7cae43334062a82cb164f2ef255013dafa2
                      metric_population
                      4576
                      ranking_population
                      4576
                      source_flag_anomaly
                      false
                      source model id
                      microsoft/phi-4
                      Original result ↗

                      Open LLM Leaderboard v2

                      52.428%

                      source id
                      openLLMLeaderboardV2
                      leaderboard name
                      Open LLM v2 BBH · Normalized
                      dimension
                      reasoning
                      metric
                      Normalized score
                      rank
                      218
                      aggregation scope
                      suite
                      Open LLM v2
                      co2_kg
                      2.7731595658140327
                      status
                      archived
                      precision
                      bfloat16
                      task_count
                      6
                      metric_kind
                      Normalized score
                      rank_method
                      Derived competition rank over published evaluations with reported scores, including variants and source flags
                      weight_type
                      Original
                      published_at
                      2025-03-20T12:17:27Z
                      chat_template
                      true
                      suite_version
                      2
                      model_revision
                      381727a5ee103da6c1b14ecd3d39cd09832cbcf8
                      source_flagged
                      false
                      missing_metrics
                        dataset_revision
                        9c09a7cae43334062a82cb164f2ef255013dafa2
                        metric_population
                        4576
                        ranking_population
                        4576
                        source_flag_anomaly
                        false
                        source model id
                        microsoft/phi-4
                        Original result ↗

                        Open LLM Leaderboard v2

                        66.906%

                        source id
                        openLLMLeaderboardV2
                        leaderboard name
                        Open LLM v2 BBH · Raw
                        dimension
                        reasoning
                        metric
                        Raw accuracy
                        rank
                        221
                        aggregation scope
                        suite
                        Open LLM v2
                        co2_kg
                        2.7731595658140327
                        status
                        archived
                        precision
                        bfloat16
                        task_count
                        6
                        metric_kind
                        Raw accuracy
                        rank_method
                        Derived competition rank over published evaluations with reported scores, including variants and source flags
                        weight_type
                        Original
                        published_at
                        2025-03-20T12:17:27Z
                        chat_template
                        true
                        suite_version
                        2
                        model_revision
                        381727a5ee103da6c1b14ecd3d39cd09832cbcf8
                        source_flagged
                        false
                        missing_metrics
                          dataset_revision
                          9c09a7cae43334062a82cb164f2ef255013dafa2
                          metric_population
                          4576
                          ranking_population
                          4576
                          source_flag_anomaly
                          false
                          source model id
                          microsoft/phi-4
                          Original result ↗

                          LMArena

                          1,158.538

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Japanese
                          dimension
                          general
                          metric
                          Arena rating
                          rank
                          234
                          sample size
                          500
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1131.8821917488722
                          confidence upper
                          1185.1945951239468
                          Original result ↗

                          LMArena

                          1,193.934

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Style Control Japanese
                          dimension
                          style
                          metric
                          Arena rating
                          rank
                          242
                          sample size
                          500
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1166.906980277133
                          confidence upper
                          1220.9604327539987
                          Original result ↗

                          LMArena

                          1,151.01

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Korean
                          dimension
                          general
                          metric
                          Arena rating
                          rank
                          244
                          sample size
                          319
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1117.898366061616
                          confidence upper
                          1184.1216172650338
                          Original result ↗

                          LMArena

                          1,222.266

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text German
                          dimension
                          general
                          metric
                          Arena rating
                          rank
                          251
                          sample size
                          523
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1196.858538014945
                          confidence upper
                          1247.6730825304596
                          Original result ↗

                          LMArena

                          1,196.347

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Style Control Korean
                          dimension
                          style
                          metric
                          Arena rating
                          rank
                          251
                          sample size
                          319
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1162.4535581665718
                          confidence upper
                          1230.2406947808295
                          Original result ↗

                          LMArena

                          1,223.884

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text French
                          dimension
                          general
                          metric
                          Arena rating
                          rank
                          253
                          sample size
                          205
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1184.2419439917744
                          confidence upper
                          1263.5262441177204
                          Original result ↗

                          LMArena

                          1,234.085

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Spanish
                          dimension
                          general
                          metric
                          Arena rating
                          rank
                          254
                          sample size
                          127
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1182.736582139311
                          confidence upper
                          1285.4332697621448
                          Original result ↗

                          LMArena

                          1,255.929

                          source id
                          lmarena
                          leaderboard name
                          LMArena Text Style Control German
                          dimension
                          style
                          metric
                          Arena rating
                          rank
                          257
                          sample size
                          523
                          evaluated at
                          2026-11-09T00:00:00.000Z
                          source model id
                          Phi-4
                          confidence lower
                          1230.0295324726762
                          confidence upper
                          1281.827816138828
                          Original result ↗

                          Published prices

                          1–4 / 4

                          microsoft/phi-4 ↗

                          $0.07

                          Inference provider
                          deepInfra
                          Platform
                          DeepInfra
                          Unit
                          input_tokens_per_million
                          Currency
                          USD
                          Billing conditions

                          microsoft/phi-4 ↗

                          $0.14

                          Inference provider
                          deepInfra
                          Platform
                          DeepInfra
                          Unit
                          output_tokens_per_million
                          Currency
                          USD
                          Billing conditions

                          microsoft/phi-4 ↗

                          $0.07

                          Inference provider
                          openRouter
                          Platform
                          OpenRouter
                          Unit
                          input_tokens_per_million
                          Currency
                          USD
                          Billing conditions

                          microsoft/phi-4 ↗

                          $0.14

                          Inference provider
                          openRouter
                          Platform
                          OpenRouter
                          Unit
                          output_tokens_per_million
                          Currency
                          USD
                          Billing conditions

                          Provider specifications & access

                          1–7 / 7

                          microsoft/phi-4 ↗

                          source model id
                          azure-huggingface/microsoft--phi-4@1
                          hosting
                          foundry-managed-compute
                          platforms
                          • Microsoft Foundry
                          license
                          MIT
                          context length
                          16384
                          architecture
                          input_modalities
                          • text
                          output_modalities
                          • text
                          capabilities

                            microsoft/phi-4 ↗

                            source model id
                            azure-huggingface/microsoft--phi-4@2
                            hosting
                            foundry-managed-compute
                            platforms
                            • Microsoft Foundry
                            license
                            MIT
                            context length
                            16384
                            architecture
                            input_modalities
                            • text
                            output_modalities
                            • text
                            capabilities

                              microsoft/phi-4 ↗

                              source model id
                              azure-huggingface/microsoft--phi-4@3
                              hosting
                              foundry-managed-compute
                              platforms
                              • Microsoft Foundry
                              license
                              MIT
                              context length
                              16384
                              architecture
                              input_modalities
                              • text
                              output_modalities
                              • text
                              capabilities

                                microsoft/phi-4 ↗

                                source model id
                                azure-huggingface/microsoft--phi-4@4
                                hosting
                                foundry-managed-compute
                                platforms
                                • Microsoft Foundry
                                license
                                MIT
                                context length
                                16384
                                architecture
                                input_modalities
                                • text
                                output_modalities
                                • text
                                capabilities

                                  microsoft/phi-4 ↗

                                  source model id
                                  azure-huggingface/microsoft--phi-4@5
                                  hosting
                                  foundry-managed-compute
                                  platforms
                                  • Microsoft Foundry
                                  license
                                  MIT
                                  context length
                                  16384
                                  architecture
                                  input_modalities
                                  • text
                                  output_modalities
                                  • text
                                  capabilities

                                    microsoft/phi-4 ↗

                                    source model id
                                    publishers/hf-microsoft/models/phi-4@001
                                    hosting
                                    vertex-self-deployed
                                    platforms
                                    • Google Cloud Vertex AI
                                    architecture
                                    input_modalities
                                    • text
                                    output_modalities
                                    • text
                                    capabilities

                                      Phi-4 ↗

                                      source model id
                                      publishers/microsoft/models/phi4@phi-4
                                      hosting
                                      vertex-self-deployed
                                      platforms
                                      • Google Cloud Vertex AI
                                      architecture
                                      input_modalities
                                      • text
                                      output_modalities
                                      • text
                                      capabilities

                                        Source specification observations

                                        1–8 / 8

                                        azureFoundry ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        azureFoundry
                                        confidence
                                        1

                                        azureFoundry ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        azureFoundry
                                        confidence
                                        1

                                        azureFoundry ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        azureFoundry
                                        confidence
                                        1

                                        azureFoundry ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        azureFoundry
                                        confidence
                                        1

                                        azureFoundry ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        azureFoundry
                                        confidence
                                        1

                                        openRouter ↗

                                        field
                                        context_length
                                        value
                                        16384
                                        normalized value
                                        16384
                                        scope
                                        source-model
                                        source id
                                        openRouter
                                        confidence
                                        1

                                        openRouter ↗

                                        field
                                        knowledge_cutoff
                                        reported field
                                        knowledge_cutoff
                                        value
                                        2024-06-30
                                        normalized value
                                        2024-06-30
                                        precision
                                        day
                                        scope
                                        source-model
                                        source id
                                        openRouter
                                        confidence
                                        1

                                        openRouter ↗

                                        field
                                        max_output_tokens
                                        value
                                        14745
                                        normalized value
                                        14745
                                        scope
                                        source-model
                                        source id
                                        openRouter
                                        confidence
                                        1

                                        Original sources

                                        1–20 / 22

                                        Hugging Face · Model repository ↗

                                        Hugging Face · microsoft/Phi-4 ↗

                                        DeepInfra · Model page ↗

                                        Google Cloud · Model page ↗

                                        Google Cloud · Model page ↗

                                        Microsoft Foundry · Model page ↗

                                        Microsoft Foundry · Model page ↗

                                        Microsoft Foundry · Model page ↗

                                        Microsoft Foundry · Model page ↗

                                        Microsoft Foundry · Model page ↗

                                        OpenRouter · Model page ↗

                                        Microsoft Foundry · Model documentation ↗

                                        arXiv · Model reference ↗

                                        DeepInfra · Catalog ↗

                                        Epoch AI · Catalog ↗

                                        GitHub · Catalog ↗

                                        Google Cloud · Catalog ↗

                                        Hugging Face · Catalog ↗

                                        Hugging Face · Catalog ↗

                                        LMArena · Catalog ↗

                                        Source links

                                        1–11 / 11

                                        Explore more models

                                        Models

                                        Loading
                                        EVIDENCE / COMPARISON

                                        Benchmarks

                                        Explore every benchmark

                                        Loading

                                        These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

                                        Install TensorTrace