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.

SpecificationOlmo 3 32B ThinkAllen AIClaude Opus 5Anthropic
ReleasedNov 21, 2025Jul 24, 2026
Context window66K1M
Max output66K128K
Input / 1M$0.15$5
Output / 1M$0.5$25
Cost per task$0.05$2.25
Arena Elo—1,393
Serving providers05
Parameters32.2BUndisclosed
Licenceapache-2.0Proprietary
Capabilities
  • Reasoning
  • 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

  • Olmo 3 32B Think
  • Claude Opus 5

  • Intelligence Index

    Olmo 3 32B Think
    6.1
    Claude Opus 5
    63.1

  • Coding Index

    Olmo 3 32B Thinknot measured
    Claude Opus 5
    78.0

  • Agentic Index

    Olmo 3 32B Thinknot measured
    Claude Opus 5
    59.2

  • Output speed

    Olmo 3 32B Thinknot measured
    Claude Opus 5
    87 t/s

  • Context window

    Olmo 3 32B Think
    66K
    Claude Opus 5
    1M

  • Latency · lower is better

    Olmo 3 32B Thinknot measured
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    Olmo 3 32B Think
    $0.237
    Claude Opus 5
    $10

  • Cost per task · lower is better

    Olmo 3 32B Think
    $0.05
    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
MetricOlmo 3 32B ThinkClaude Opus 5
Intelligence Index6.163.1
Coding Index—78.0
Agentic Index—59.2
Output speed—87 t/s
Context window66K1M
Latency · lower is better—1.12s
Blended price / 1M · lower is better$0.237$10
Cost per task · lower is better$0.05$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • Olmo 3 32B Think
  • Claude Opus 5

  • GPQA Diamond

    Olmo 3 32B Think
    61.0%
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    Olmo 3 32B Think
    6.4%
    Claude Opus 5
    54.9%

  • SciCode

    Olmo 3 32B Think
    28.6%
    Claude Opus 5
    55.7%

  • τ²-bench

    Olmo 3 32B Think
    0.0%
    Claude Opus 5
    42.1%

  • Terminal-Bench Hard

    Olmo 3 32B Think
    1.5%
    Claude Opus 5not measured

  • LiveCodeBench

    Olmo 3 32B Think
    67.2%
    Claude Opus 5not measured

  • AA-LCR long context

    Olmo 3 32B Think
    0.0%
    Claude Opus 5
    75.7%

  • AIME 2025

    Olmo 3 32B Think
    73.7%
    Claude Opus 5not measured

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

View as table
EvaluationOlmo 3 32B ThinkClaude Opus 5
GPQA Diamond61.0%93.2%
Humanity's Last Exam6.4%54.9%
SciCode28.6%55.7%
τ²-bench0.0%42.1%
Terminal-Bench Hard1.5%—
LiveCodeBench67.2%—
AA-LCR long context0.0%75.7%
AIME 202573.7%—

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