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.

SpecificationQwen3 VL 8B InstructQwenClaude Opus 5Anthropic
ReleasedOct 14, 2025Jul 24, 2026
Context window262K1M
Max output33K128K
Input / 1M$0.117$5
Output / 1M$0.455$25
Cost per task$0.02$2.25
Arena Elo1,393
Serving providers25
Parameters8.8BUndisclosed
Licenceapache-2.0Proprietary
Capabilities
  • Tool use
  • Vision
  • 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

  • Qwen3 VL 8B Instruct
  • Claude Opus 5
  • Intelligence Index

    Qwen3 VL 8B Instruct
    8.2
    Claude Opus 5
    63.1
  • Coding Index

    Qwen3 VL 8B Instructnot measured
    Claude Opus 5
    78.0
  • Agentic Index

    Qwen3 VL 8B Instructnot measured
    Claude Opus 5
    59.2
  • Output speed

    Qwen3 VL 8B Instruct
    51 t/s
    Claude Opus 5
    87 t/s
  • Context window

    Qwen3 VL 8B Instruct
    262K
    Claude Opus 5
    1M
  • Latency · lower is better

    Qwen3 VL 8B Instruct
    357ms
    Claude Opus 5
    1.12s
  • Blended price / 1M · lower is better

    Qwen3 VL 8B Instruct
    $0.202
    Claude Opus 5
    $10
  • Cost per task · lower is better

    Qwen3 VL 8B Instruct
    $0.02
    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
MetricQwen3 VL 8B InstructClaude Opus 5
Intelligence Index8.263.1
Coding Index78.0
Agentic Index59.2
Output speed51 t/s87 t/s
Context window262K1M
Latency · lower is better357ms1.12s
Blended price / 1M · lower is better$0.202$10
Cost per task · lower is better$0.02$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • Qwen3 VL 8B Instruct
  • Claude Opus 5
  • GPQA Diamond

    Qwen3 VL 8B Instruct
    42.7%
    Claude Opus 5
    93.2%
  • Humanity's Last Exam

    Qwen3 VL 8B Instruct
    2.7%
    Claude Opus 5
    54.9%
  • SciCode

    Qwen3 VL 8B Instruct
    17.4%
    Claude Opus 5
    55.7%
  • τ²-bench

    Qwen3 VL 8B Instruct
    29.2%
    Claude Opus 5
    42.1%
  • Terminal-Bench Hard

    Qwen3 VL 8B Instruct
    2.3%
    Claude Opus 5not measured
  • LiveCodeBench

    Qwen3 VL 8B Instruct
    33.2%
    Claude Opus 5not measured
  • AA-LCR long context

    Qwen3 VL 8B Instruct
    16.7%
    Claude Opus 5
    75.7%
  • AIME 2025

    Qwen3 VL 8B Instruct
    27.3%
    Claude Opus 5not measured

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

View as table
EvaluationQwen3 VL 8B InstructClaude Opus 5
GPQA Diamond42.7%93.2%
Humanity's Last Exam2.7%54.9%
SciCode17.4%55.7%
τ²-bench29.2%42.1%
Terminal-Bench Hard2.3%
LiveCodeBench33.2%
AA-LCR long context16.7%75.7%
AIME 202527.3%

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