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.

SpecificationStep 3.5 FlashStepFunClaude Opus 5Anthropic
ReleasedJan 29, 2026Jul 24, 2026
Context window262K1M
Max output66K128K
Input / 1M$0.1$5
Output / 1M$0.3$25
Cost per task$0.03$2.25
Arena Elo1,393
Serving providers15
Parameters199.4BUndisclosed
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

  • Step 3.5 Flash
  • Claude Opus 5
  • Intelligence Index

    Step 3.5 Flash
    26.5
    Claude Opus 5
    63.1
  • Coding Index

    Step 3.5 Flashnot measured
    Claude Opus 5
    78.0
  • Agentic Index

    Step 3.5 Flashnot measured
    Claude Opus 5
    59.2
  • Output speed

    Step 3.5 Flash
    56 t/s
    Claude Opus 5
    87 t/s
  • Context window

    Step 3.5 Flash
    262K
    Claude Opus 5
    1M
  • Latency · lower is better

    Step 3.5 Flash
    1.39s
    Claude Opus 5
    1.12s
  • Blended price / 1M · lower is better

    Step 3.5 Flash
    $0.15
    Claude Opus 5
    $10
  • Cost per task · lower is better

    Step 3.5 Flash
    $0.03
    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
MetricStep 3.5 FlashClaude Opus 5
Intelligence Index26.563.1
Coding Index78.0
Agentic Index59.2
Output speed56 t/s87 t/s
Context window262K1M
Latency · lower is better1.39s1.12s
Blended price / 1M · lower is better$0.15$10
Cost per task · lower is better$0.03$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • Step 3.5 Flash
  • Claude Opus 5
  • GPQA Diamond

    Step 3.5 Flash
    82.6%
    Claude Opus 5
    93.2%
  • Humanity's Last Exam

    Step 3.5 Flash
    24.5%
    Claude Opus 5
    54.9%
  • SciCode

    Step 3.5 Flash
    38.5%
    Claude Opus 5
    55.7%
  • τ²-bench

    Step 3.5 Flash
    87.4%
    Claude Opus 5
    42.1%
  • Terminal-Bench Hard

    Step 3.5 Flash
    32.6%
    Claude Opus 5not measured
  • AA-LCR long context

    Step 3.5 Flash
    60.3%
    Claude Opus 5
    75.7%

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

View as table
EvaluationStep 3.5 FlashClaude Opus 5
GPQA Diamond82.6%93.2%
Humanity's Last Exam24.5%54.9%
SciCode38.5%55.7%
τ²-bench87.4%42.1%
Terminal-Bench Hard32.6%
LiveCodeBench
AA-LCR long context60.3%75.7%
AIME 2025

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