GPT-5.1
Faster, Warmer GPT-5 Iteration
GPT-5.1 was OpenAI’s first “fast follow” iteration on GPT-5: less token-heavy thinking on easy tasks, smoother day‑to‑day chat, and better instruction following. It kept the same “one surface with adaptive reasoning” product idea, but tuned the tradeoff between speed and depth and added a developer-facing no‑reasoning option for latency.
Intelligence
Top 10%
Speed
Medium
131 tok/s output
Cost
High
$1.25 in / $10.00 out
Context
272K
Up to 272,000 tokens
How are Intelligence, Speed & Cost bucketed?
Intelligence and Speed buckets are percentile ranks on
Artificial Analysis. Cost buckets are fixed dollar
thresholds keyed off output-token price ($/M out).
Intelligence
- Top 1%≤ 1%
- Top 5%≤ 5%
- Top 10%≤ 10%
- Good≤ 25%
- Medium≤ 50%
- Below avg> 50%
Speed
- Top 1%≥ 345 tok/s
- Top 5%≥ 237 tok/s
- Top 10%≥ 196 tok/s
- Good≥ 146 tok/s
- Medium≥ 90 tok/s
- Slow< 90 tok/s
Cost
- Freeopen weights · self-host
- Low< $1 / M out
- Moderate$1–5 / M out
- High≥ $5 / M out
Why it matters
It normalized versioned GPT-5 generations (5.1 → 5.2 → 5.3/5.4/5.5) and set expectations for rapid product iteration in the same model line.
Core Capabilities
Long Documents
Handles entire codebases, books, and multi-doc RAG.
Multimodal
Combines text, vision, and audio in one model.
Generative
Produces images, video, audio, or other media.
Reasoning
Solves complex math, logic, and planning tasks.
Context Window
272k tokens
≈ 209+ pages
4k Chat 聊天
32k Long docs 长文档
128k Books 整本书
400k Multi-doc 多文档
1M Codebase 整个代码库
10M
272k
Availability
API
Available
Product / App
Not available
Open Source
Not released
Enterprise
Contact sales
Pricing Model
Pay per token
Input and output billed separately.
Pay-per-token Capability / Performance
Where this model sits relative to the middle 60% of models in the tree. All scores are 0–10 (higher is better).
Lower 20% Upper 80% This model
Context / memory
Context window size · log-scaled
6.0
9.0
7.1
Lower 20% 20th percentile — 20% of models score below this This model Where the current model lands Upper 80% 80th percentile — only 20% of models score above this
Percentile boundaries are computed across every model in the tree that reports the underlying benchmark for each capability.
What it feels like
- Adaptive thinking — auto-tunes reasoning depth per task; not a fixed-tier release
- 2-3x faster than GPT-5 on tool-heavy reasoning at similar or better quality (Balyasny eval)
- Uses ~half the tokens of leading competitors on agent flows for the same answer
- CodeRabbit: 'top model of choice for PR reviews' at release
- Cognition: 'noticeably better at understanding what you're asking and working with you to get it done'
- 'No reasoning' mode for fast everyday answers; you don't pay for thinking when it's not needed
Best use cases
- Production agent workflows where token cost / latency dominate
- PR review and code-discussion bots (CodeRabbit picked it as default)
- Cursor / Devin-style developer tools where adaptive depth matters
- Replacing GPT-5 in production where speed is the constraint
Tools to try
Not ideal for
- Hardest research-grade reasoning — GPT-5.2 Thinking and Pro are stronger
- Agentic-coding leaderboards on raw SWE-bench — Claude Opus 4.5 / 4.6 still lead
- Workloads where you want manual control over reasoning effort tier
Model Evolution
GPT is OpenAI's language model family.