Step 3.5 Flash vs GPT-6 Luna

Put models side by side on price, context, capabilities, and benchmarks.

Step 3.5 Flash is StepFun's flagship sparse MoE reasoning model (196B parameters, 11B active) built for fast, reliable execution of complex tasks. It excels at decomposing and planning multi-step problems and orchestrating tool calls, making it well suited for agentic and coding workflows.

Learn more
AuthorStepfun
CountryπŸ‡¨πŸ‡³ China
ReleasedJan 29, 2026
Overview
Cost rate
0.1x
Context
262K
Input
Output
Reasoning
Tool calling
Structured outputs
Pricing/M tokens
Input$0.10
Cached input$0.00
Output$0.30
Context
Context length262,144
Max output tokens65,536
Knowledge cutoff-
Performance
Throughput156 tok/s
Latency3.45s

GPT-6 Luna is the fast, low-cost tier of OpenAI's GPT-6 series (1.05M-token context, text, image, and file input, $0.10/$0.50 per million input/output tokens), released September 22, 2026 below GPT-6 Sol and the flagship GPT-6 Astra. It is built for high-volume tasks with a clear goal, such as summarization, extraction, classification, quick questions, and lightweight agentic steps.

Image understanding#68Mathematical#74Software & IT Services#79Writing, Literature, & Language#93
Learn more
AuthorOpenAI
CountryπŸ‡ΊπŸ‡Έ United States
ReleasedSep 22, 2026
Overview
Cost rate
0.1x
Context
1.1M
Input
Output
Reasoning
Tool calling
Structured outputs
Pricing/M tokens
Input$0.10
Cached input$0.01
Output$0.50
Context
Context length1,050,000
Max output tokens128,000
Knowledge cutoff-
Performance
Throughput93 tok/s
Latency5.62s

Benchmarks

Head-to-head reasoning, coding and agentic scores, plus per-domain rankings.

Intelligence

64.6
Claude Fable 5.2 (Max, Default Fallback)
62.1
GPT-6.2 Astra (Max)
61.4
Gemini 4 Neon (Xhigh)
60.0
Gemini 4 Carbon (High)
59.1
Claude Opus 5.5 (Max, Default Fallback)
23.5
GPT-6 Luna

Step 3.5 Flash has no data

Source: Artificial Analysis

Coding

1824
Claude Fable 5.2 (xHigh)
1818
Claude Opus 5.5 (Max)
1781
GPT-6.2 Astra (Max)
1773
Claude Sonnet 5.5 (xHigh)
1769
Gemini 4 Carbon (High)
1578
GPT-6 Luna

Step 3.5 Flash has no data

Source: Code Arena

Automation & Tool Use

0.27
Claude Fable 5.2 (Max)
0.19
Claude Opus 5.5 (Max)
0.18
Gemini 4 Neon (High)
0.16
Gemini 4 Carbon (High)
0.16
Claude Sonnet 5.5 (Max)
0.01
GPT-6 Luna

Step 3.5 Flash has no data

Source: Agent Arena

Software & IT Services

arena.ai
1548
Gemini 4 Carbon (High)
1548
Gemini 4 Argon (High)
1546
Gemini 4 Neon (High)
1545
Muse Wonder (Max)
1540
Claude Opus 4.7 (High)
1538
Claude Fable 5 (High)
1537
Claude Opus 4.6 (High)
1535
Claude Opus 5.5 (High)
1531
GPT-6.1 Sol (Max)
1490
GPT-6 Luna

Step 3.5 Flash has no data

Popular comparisons