Skip to content
llmwaves

Compare

Four models, one page

Pick the models you are actually choosing between. Every row is scaled within itself, so a bar means something next to its neighbours and nothing across rows.

Specificationgpt-oss-20bOpenAIClaude Opus 5Anthropic
ReleasedAug 5, 2025Jul 24, 2026
Context window131K1M
Max output131K128K
Input / 1M$0.03$5
Output / 1M$0.13$25
Cost per task$0.01$2.25
Arena Elo9611,393
Serving providers115
Parameters21.5BUndisclosed
Licenceapache-2.0Proprietary
Capabilities
  • Reasoning
  • Tool use
  • Open weights
  • Reasoning
  • Tool use
  • Vision

Metrics side by side

Each row is scaled to the largest value in that row — bars compare within a row, never across rows

  • gpt-oss-20b
  • Claude Opus 5
  • Intelligence Index

    gpt-oss-20b
    15.2
    Claude Opus 5
    63.1
  • Coding Index

    gpt-oss-20b
    20.7
    Claude Opus 5
    78.0
  • Agentic Index

    gpt-oss-20b
    3.1
    Claude Opus 5
    59.2
  • Output speed

    gpt-oss-20b
    252 t/s
    Claude Opus 5
    87 t/s
  • Context window

    gpt-oss-20b
    131K
    Claude Opus 5
    1M
  • Latency · lower is better

    gpt-oss-20b
    259ms
    Claude Opus 5
    1.12s
  • Blended price / 1M · lower is better

    gpt-oss-20b
    $0.055
    Claude Opus 5
    $10
  • Cost per task · lower is better

    gpt-oss-20b
    $0.01
    Claude Opus 5
    $2.25

Rows marked “lower is better” still draw a longer bar for a larger number — read the value, not just the length. Arena Elo is in the specification table above instead: it has no meaningful zero, so a bar would flatten the gaps.

View as table
Metricgpt-oss-20bClaude Opus 5
Intelligence Index15.263.1
Coding Index20.778.0
Agentic Index3.159.2
Output speed252 t/s87 t/s
Context window131K1M
Latency · lower is better259ms1.12s
Blended price / 1M · lower is better$0.055$10
Cost per task · lower is better$0.01$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • gpt-oss-20b
  • Claude Opus 5
  • GPQA Diamond

    gpt-oss-20b
    68.8%
    Claude Opus 5
    93.2%
  • Humanity's Last Exam

    gpt-oss-20b
    11.0%
    Claude Opus 5
    54.9%
  • SciCode

    gpt-oss-20b
    34.4%
    Claude Opus 5
    55.7%
  • τ²-bench

    gpt-oss-20b
    60.2%
    Claude Opus 5
    42.1%
  • Terminal-Bench Hard

    gpt-oss-20b
    10.6%
    Claude Opus 5not measured
  • LiveCodeBench

    gpt-oss-20b
    77.7%
    Claude Opus 5not measured
  • AA-LCR long context

    gpt-oss-20b
    33.3%
    Claude Opus 5
    75.7%
  • AIME 2025

    gpt-oss-20b
    89.3%
    Claude Opus 5not measured

A missing bar means that evaluation was not run for that model — it is not a zero.

View as table
Evaluationgpt-oss-20bClaude Opus 5
GPQA Diamond68.8%93.2%
Humanity's Last Exam11.0%54.9%
SciCode34.4%55.7%
τ²-bench60.2%42.1%
Terminal-Bench Hard10.6%
LiveCodeBench77.7%
AA-LCR long context33.3%75.7%
AIME 202589.3%

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).