RankingIA v0.6.1
EN

Step 5 Previewopen

StepFun

56.7#95 of 427 overall
Input / output
US$ 1,00 / US$ 2,70
US$ per 1M tokens
Context
1.0M
tokens
Release
10/08/2026
release date
Speed
163 tok/s
first token in 0.70 s

Family history

The 2 versions of this family, from first to newest, by release date. Click a point to open that version.

See the data as a table
ModelReleasedScoreRankAA Intel.Elo Arena
Step-3————1350
Step 5 Preview10/08/202656.7#9528.51443

Scores by source

Scores by sourcepercentile = % of ranked models it beats
MetricValuePercentileSource
LiveBench weight 3 58.6 26% LiveBench
AA Intelligence Index weight 3 28.5 51% Artificial Analysis
Arena text (Elo) weight 2 1443 73% LMArena
Arena WebDev (Elo) weight 1 1571 70% LMArena
Arena vision (Elo) weight 0.5 1265 64% LMArena
Humanity's Last Exam weight 1 46.5 95% Artificial Analysis
Terminal-Bench weight 1 33.3 86% Artificial Analysis
Long-context reasoning 88.3 100% Artificial Analysis
LiveBench long agentic 19.4 29% LiveBench
LiveBench reasoning 83.8 42% LiveBench
LiveBench coding 85.6 52% LiveBench
LiveBench browser 24.8 23% LiveBench
LiveBench data analysis 63.5 16% LiveBench
LiveBench research 72.8 42% LiveBench
LiveBench design 76.9 48% LiveBench

By category

CategoryScoreRankMetrics
Agentic54.9#52LiveBench long agentic, Terminal-Bench
Coding57.1#35LiveBench coding, Arena WebDev (Elo)
Reasoning62.4#128LiveBench reasoning, Humanity's Last Exam
Data analysis33.0#27LiveBench data analysis
Research64.0#71LiveBench research, Long-context reasoning
Design56.1#39LiveBench design, Arena WebDev (Elo)
Browser use36.3#25LiveBench browser
Chat61.4#54Arena text (Elo)
Vision56.9#39Arena vision (Elo)

History

RankingIA score on each collection day (hover to see the rank).

Daily collection started on 10/30/2026. The chart appears from the second day the model is ranked.

Compare with other models

At the sources