AI Models Navi
Back to Models
OpenAIProprietary

GPT-5.4

GPT-5.4 is the latest multimodal reasoning model in the GPT-5 series developed by OpenAI. It is optimized for complex knowledge work, software engineering, and long-context analysis up to 1 million tokens.

Parameters

Undisclosed

Context Window

License

Proprietary

Release Date

2026-03-05

API Pricing

Input Price (per 1M tokens)

$2.5

Output Price (per 1M tokens)

$

Billing Mode: standard

Strengths

  • ・Advanced software development capabilities
  • ・Long-context analysis up to 1 million tokens
  • ・Industry-leading on key evaluation benchmarks

Weaknesses

  • ・Model weights are not public
  • ・Closed-source license system
  • ・Incurs API usage costs

Use Cases

  • ・Advanced software engineering support
  • ・Business tasks with complex specialized knowledge
  • ・Ultra-long context analysis

Deep Analysis

Arena Elo

1467

#9 overall

SWE-Bench Verified

80.8%

vs Claude Opus 4.7: 87.6%

SWE-Bench Pro

57.7%

vs Claude Opus 4.7: 64.3%

GPQA Diamond

94.4%

near parity with frontier

Input Price

$2.50/1M

vs Claude Opus 4.7: $5/1M

Context Window

1.05M tokens

largest in GPT-5 family

Strengths

  • ・Best-in-class web research (BrowseComp 89.3%)
  • ・Strong terminal-heavy coding (Terminal-Bench 75.1%)
  • ・Competitive pricing at $2.50/$15 per 1M tokens
  • ・1M context window with strong retrieval

Weaknesses

  • ・Trails Claude Opus 4.7 on agentic coding (SWE-bench Pro 57.7% vs 64.3%)
  • ・Weaker multi-tool orchestration than Claude (MCP-Atlas 68.1% vs 77.3%)
  • ・Computer use below Claude Opus 4.7 (OSWorld 75.0% vs 78.0%)

Competitor Comparison

ModelArenaSWEGPQAPrice
Claude Opus 4.7~150087.6%94.2%$5/$25
Gemini 3.1 Pro~148584.2%94.3%$2/$12
GPT-5.4146780.8%94.4%$2.50/$15

Sources

Analysis generated: 2026-05-24