Recent studyBlind TTS Elo is live. Compare two anonymous voice samples, vote after listening, and help separate real preference signal from noise.Vote in the study ->
Codesota · Models · Step-3.5-Flash Base11 results · 11 benchmarks
Model card

Step-3.5-Flash Base.

unknown
§ 02 · Benchmarks

Every benchmark Step-3.5-Flash Base has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01SimpleQANatural Language Processing · Question Answeringaccuracy31.6%#2/7source ↗
02HellaSwagReasoning · Commonsense Reasoningaccuracy90.2%#3/17source ↗
03MBPP+Computer Code · Code Generationpass-179.4%#3/9source ↗
04BIG-Bench HardReasoning · Multi-step Reasoningaccuracy88.2%#4/11source ↗
05HumanEval+Computer Code · Code Generationpass-181.1%#5/12source ↗
06WinoGrandeReasoning · Commonsense Reasoningaccuracy79.1%#5/13source ↗
07GSM8KReasoning · Mathematical Reasoningaccuracy88.2%#36/48source ↗
08MATHReasoning · Mathematical Reasoningaccuracy66.8%#39/46source ↗
09MMLUReasoning · Commonsense Reasoningaccuracy85.8%#44/64source ↗
10MMLU-ProReasoning · Commonsense Reasoningaccuracy62.3%#69/73source ↗
11GPQA DiamondReasoning · Multi-step Reasoningaccuracy41.7%#71/74source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area

Where Step-3.5-Flash Base actually performs.

Natural Language Processing
1
benchmark
avg rank #2.0
Computer Code
2
benchmarks
avg rank #4.0
Reasoning
8
benchmarks
avg rank #33.9
§ 04 · Papers

1 paper with results for Step-3.5-Flash Base.

  1. 2026-02-11· 11 results

    Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters

§ 06 · Sources & freshness

Where these numbers come from.

pwc-dump
11
results
0 of 11 rows marked verified.