Claude Sonnet 4.6
Anthropic · released Feb 17, 2026
Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work. It excels at iterative development, complex codebase navigation, end-to-end project management with...
Specification
- Context window
- 1M
- Max output
- 128K
- Knowledge cutoff
- Not stated
- Parameters
- Undisclosed
- Licence
- Proprietary
- Serving providers
- 4
- Moderated
- Yes
- Uptime
- 100.0%
Intelligence
36.8
84th percentile
Coding
—
Coding Index
Agentic
42.1
Agentic Index
Output speed
52 t/s
Median across providers
Latency
843ms
Time to first token
Cost per task
$1.35
Estimated
Benchmarks
Where the score comes from
The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.
Evaluation scores
Percentage correct · higher is better
- GPQA Diamond79.9%
- τ²-bench (Telecom)79.5%
- AA-LCR (long context)62.3%
- SciCode46.9%
- Terminal-Bench Hard46.2%
- IFBench41.2%
- Humanity's Last Exam13.3%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| GPQA Diamond | 79.9% |
| τ²-bench (Telecom) | 79.5% |
| AA-LCR (long context) | 62.3% |
| SciCode | 46.9% |
| Terminal-Bench Hard | 46.2% |
| IFBench | 41.2% |
| Humanity's Last Exam | 13.3% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- Grok 4.3xAI37.9
- Ling-3.0-flashInclusionAI37.8
- Qwen3.6 27BQwen37.7
- GPT-5.1OpenAI37.5
- Gemini 3.5 Flash LiteGoogle37.4
- Claude Sonnet 4.6Anthropic36.8
- Kimi K2.5Moonshot AI36.0
- Claude Opus 4.5Anthropic35.6
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| Grok 4.3 | 37.9 |
| Ling-3.0-flash | 37.8 |
| Qwen3.6 27B | 37.7 |
| GPT-5.1 | 37.5 |
| Gemini 3.5 Flash Lite | 37.4 |
| Claude Sonnet 4.6 | 36.8 |
| Kimi K2.5 | 36.0 |
| Claude Opus 4.5 | 35.6 |
Percentile among all indexed models
Pricing
What it costs to run
List prices per million tokens, plus what one representative task works out to.
List price
- Input / 1M tokens
- $3
- Output / 1M tokens
- $15
- Cached input / 1M
- $0.3
- Blended 3:1
- $6
One task, estimated
$1.35
- Input tokens
- 50,000
- Output tokens
- 80,000
- Profile
- Reasoning
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).
Serving providers
Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.
Arena
Head-to-head generation quality
Elo from pairwise judgements, broken out by the kind of thing the model was asked to build.
- Overall Elo
- 1,308
- Win rate
- 60.1%
- Strongest at
- Websites
- Tournaments
- 13,607
Elo by category
Dot position on a 1,225–1,325 scale · Elo has no meaningful zero
- 1,308
- 1,308
- 1,308
- 1,301
- 1,290
- 1,244
- 1,271
- 1,252
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| Websites | 1308 | 60.1% |
| UI components | 1308 | 58.4% |
| Data visualisation | 1308 | 58.2% |
| Game development | 1301 | 59.0% |
| 3D scenes | 1290 | 57.6% |
| SVG | 1244 | 58.8% |
| Mobile apps | 1271 | 63.4% |
| Full-stack apps | 1252 | 64.1% |