step-5-preview-high: benchmark scores, price and ranking

step-5-preview-high from StepFun ranks #219 of 743 LLMs on the LLMs Tiger aggregate benchmark score (50.0), based on 2 benchmark results captured up to 2026-10-13.

Organisation: StepFun
Type: Preview release
Open weights: No
Cheapest API price: $1.00 in / $2.70 out per million tokens (via openrouter, observed 2026-10-13T20:47Z)
Aggregate rank: #219 of 743 (score 50.0)

Across the 2 benchmarks where at least ten models are scored, step-5-preview-high is in the top 10% on 0 and in the top half on 1.

Its cheapest listed input price, $1.00 per million tokens, is lower than that of 35% of the 427 models with a listed price.

step-5-preview-high benchmark results

step-5-preview-high benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
WebDev Arena	1,570	#16 of 67	Epoch AI Benchmarking Hub	2026-10-13
LMArena Code Elo	1,570	#23 of 30	LMArena	2026-10-06

About step-5-preview-high

step-5-preview-high is a preview release from StepFun. The cheapest API listing we track charges $1.00 per million input tokens and $2.70 per million output tokens, through openrouter. We hold 2 benchmark results for it across 1 area, and none is in the top tenth of its board.

Where step-5-preview-high stands

On our aggregate score, step-5-preview-high is #219 of 743 ranked models, around the middle of the table. Its score is 50.0; the scale is centred on 50, so it sits 0.0 points above the typical model. That beats 71% of the ranked field. Among the 2 ranked StepFun models it is #1. The next one down is Step 3.7 Flash at 47.0.

step-5-preview-high, benchmark by benchmark

WebDev Arena puts step-5-preview-high at 1,570, #16 of 67. That beats 76% of the models tested, but 250 points short of Claude Opus 5.5 (max) (1,820). The results either side are GPT-6 Luna (max) (1,582) and Gemini 3.8 Flash (high) (1,568). It covers building web apps, crowd-voted (Elo). The board median is 1,433, so step-5-preview-high is 137 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

step-5-preview-high lands below the middle on LMArena Code Elo: 1,570, #23 of 30, with 73% of the field scoring higher. Claude Opus 5.5 (max) leads at 1,814; the closest result above is GPT-6 Luna (max) at 1,579. It covers crowd-voted coding answers. The board median is 1,622, so step-5-preview-high is 52 points below it. Source: LMArena, captured 2026-10-06.

Strengths and weaknesses

Software engineering: uneven, ahead of 50% of the field on average across 2 benchmarks. Its best board here is WebDev Arena (#16 of 67) and its weakest is LMArena Code Elo (#23 of 30).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so step-5-preview-high costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), DeepSeek V4 Flash 0731 (max) ($0.016, 53.1). Similar scores cost more on Grok 4.20 (Reasoning) ($1.25) and GPT-5 (high) ($1.25).

Reading these numbers

The 2 scores on this page come from 2 sources: LMArena (1), Epoch AI Benchmarking Hub (1). No official leaderboard supplied any of them, so treat them as third-party measurements. The leaderboard also lists 2 other reasoning-effort settings of this model, which redirect here. This page shows the highest. With 2 results the aggregate rank can move a lot when one new score arrives.

step-5-preview-high is also listed at 2 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

#216 grok-3-beta (50.0)
#217 Claude Sonnet 4.6 (max) (50.0)
#218 grok-4-20 (50.0)
#220 claude-sonnet-4-20250514_2K (50.0)
#221 gemini-exp-1114 (50.0)
#222 gemini-2.5-flash-preview-09-2025 (50.0)