Automated page. Written and translated automatically by AI and checked against the sources by software; no human editor. SI-IQ is an index built from public benchmark results, not an IQ test. How it works

How smart is Claude Sonnet 4.6?

113 SI-IQ · 90% interval 108–118
Rank 55 of 80 ranked models · Plausible rank range 36–65

Claude Sonnet 4.6 is a model from Anthropic, released on Feb 17, 2026. Its SI-IQ is 113 (90% interval 108–118), estimated from 7 benchmark results by 2 independent evaluators. Its strongest area is Reasoning and its weakest is Knowledge.

Epoch AI and OpenRouter measured different GPQA Diamond scores for this model (more than 10 points apart), probably with different settings. Both runs count, each with half weight.

Facts

Developer: Anthropic
Released: Feb 17, 2026
Access: API access
Benchmarks used: 7 results from 2 evaluators
API price: $3.00 input / $15 output per 1M tokens (developer's price list, Oct 18, 2026)
Context window: 1,000,000 tokens

Ability by area (SI-IQ, 90% interval)

Every benchmark result we used

Each score is one value exactly as the source published it: for Epoch AI the best of its run settings, for OpenRouter the median hosting provider (the developer's own hosting when it ran the test). “Implied SI-IQ” is the ability that this single result points to; the overall SI-IQ is their precision-weighted average.

Benchmark	Area	Score	Implied SI-IQ	Evaluated by	As listed by the source	Retrieved	Source
Chess Puzzles	Reasoning	23.0%	115	Epoch AI	`claude-sonnet-4-6_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
Mystery Game Puzzles	Reasoning	16.0%	107	Epoch AI	`claude-sonnet-4-6_low`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
OTIS Mock AIME 2024–2025	Math	85.8%	112	Epoch AI	`claude-sonnet-4-6_32K`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
SWE-bench Verified (Epoch AI run)	Coding	75.2%	122	Epoch AI	`claude-sonnet-4-6`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
GPQA Diamond	Knowledge	87.4%	119	Epoch AI	`claude-sonnet-4-6_32K`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
GPQA Diamond (OpenRouter run)	Knowledge	64.6%	91	OpenRouter	`anthropic/claude-4.6-sonnet-20260217` Anthropic · first-party provider (1 of 5 rows)	Oct 18, 2026	OpenRouter benchmark datasets · data file
SimpleQA Verified	Knowledge	35.5%	115	Epoch AI	`claude-sonnet-4-6_high`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
τ²-bench Airline (OpenRouter run)	Agents	74.9%	129	OpenRouter	`anthropic/claude-4.6-sonnet-20260217` Anthropic · first-party provider (1 of 5 rows)	Oct 18, 2026	OpenRouter benchmark datasets · data file

Human preference (Arena)

On the Arena leaderboard, where people vote between two anonymous answers, Claude Sonnet 4.6 has a rating of 1,472 (1,469–1,476) from 70,931 votes (published Oct 15, 2026). Human preference is a different measure from benchmark ability and is not part of SI-IQ. Arena (Hugging Face)

Models with similar SI-IQ

Model	SI-IQ
Qwen3.5 397B-A17BAlibaba	114	Claude Sonnet 4.6 vs Qwen3.5 397B-A17B
InklingThinking Machines	114	Claude Sonnet 4.6 vs Inkling
GLM-4.7Z.ai (Zhipu AI)	112	Claude Sonnet 4.6 vs GLM-4.7
Claude Opus 4.5Anthropic	115	Claude Sonnet 4.6 vs Claude Opus 4.5
Qwen3.7-PlusAlibaba	115	Claude Sonnet 4.6 vs Qwen3.7-Plus

Computed with SI-IQ method version 1.0.0. Read the full methodology

Data sources: Arena leaderboard dataset on Hugging Face (CC BY 4.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); API list prices via LiteLLM's price table (MIT); each price links to the developer's pricing page; Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..