Cheaper alternatives to Claude Sonnet 4.6

Anthropic LMArena 1457.6 $6.00/M on a balanced workload prices as of 2026-08-28

2 models are both better and cheaper, giving up nothing.

Of the 133 models carrying both an independent score and a published price at the same delivery mode, 12 score at least as high as Claude Sonnet 4.6 and cost no more under at least one workload. 11 of them are genuinely cheaper; the rest match the price and win on score alone. 2 match or beat it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. 10 do not, and every row names what it drops.

Dominance is not a property of a model. It is a property of a model, a capability lens and a workload, and all three are named on every row. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset; prices are blended per million tokens. A higher score is not a drop-in replacement.

This page is built from that comparison and nothing else. When no model is both better and cheaper than Claude Sonnet 4.6, it is not generated.

Where it holds

Workload	Claude Sonnet 4.6 $/M	Better and cheaper	With a trade
Balanced	$6.00	2	10
Summarise	$2.83	2	10
Chat	$6.99	2	10
Code gen	$9.98	2	10
Agentic	$3.19	2	10

Workload mixes are defined on the methodology page. Cached input is priced at the cached rate, and reasoning tokens at the worse of the reasoning and output rates.

What you would be replacing

Intelligence	1457.6
Context window	1M
Max output	128K
Input modes	text, image, file
Tool use	yes
Extended reasoning	yes

A replacement has to clear every line above, not just the score. Full record for Claude Sonnet 4.6.

Better and cheaper, nothing given up 2

Each of these matches or beats Claude Sonnet 4.6 on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

#	Model	Intelligence	$/M	Holds under
1	Muse Spark 1.1 Meta	1478.3 +20.7	$2.00	Balanced −67% Summarise −62% Chat −70% Code gen −70% Agentic −67%
2	GPT-5.4 OpenAI	1469.6 +12	$5.63	Balanced −6% Summarise −12% Chat −2% Code gen −2% Agentic −5%

Cheaper and higher-scoring, but you give something up 10

These score at least as high and cost no more on the two plotted axes, and lose something that is not on them. Read the last column before switching.

Model	Intelligence	$/M	Holds under	What you give up
Gemini 3.7 Flash Google	1490.2 +32.6	$0.750	Balanced −88% Summarise −88% Chat −87% Code gen −88% Agentic −87%	128K → 66K max output
Gemini 3.5 Flash Google	1482.6 +25	$3.38	Balanced −44% Summarise −47% Chat −41% Code gen −41% Agentic −43%	128K → 66K max output
Qwen3.8 Max Qwen	1481.9 +24.3	$3.00	Balanced −50% Summarise −40% Chat −56% Code gen −57% Agentic −51%	no file input
Gemini 3.6 Flash Google	1476.5 +18.9	$1.50	Balanced −75% Summarise −75% Chat −75% Code gen −75% Agentic −75%	128K → 66K max output
GLM 5.3 Z.ai	1476.9 +19.3	$2.15	Balanced −64% Summarise −57% Chat −68% Code gen −69% Agentic −63%	no image, file input
Gemini 3.1 Pro Preview Google	1479.5 +21.9	$4.50	Balanced −25% Summarise −30% Chat −22% Code gen −21% Agentic −24%	128K → 66K max output
Kimi K3 Moonshot AI	1476.3 +18.7	$6.00	Balanced same price Summarise same price Chat same price Code gen same price Agentic same price	no file input
MiMo-V2.5-Pro Xiaomi	1465 tie	$0.544	Balanced −91% Summarise −88% Chat −93% Code gen −93% Agentic −92%	no image, file input
GLM 5.2 Z.ai	1465.4 tie	$1.83	Balanced −70% Summarise −63% Chat −73% Code gen −74% Agentic −69%	no image, file input
GLM 5.1 Z.ai	1464.1 tie	$1.94	Balanced −68% Summarise −61% Chat −71% Code gen −72% Agentic −67%	1.0M → 205K context no image, file input

What this compares, and what it leaves out

Quality is LMArena Elo, used under CC BY 4.0 from the official dataset. The 95% confidence interval on a difference between two scores is about ±10.68 points, so a gap smaller than that is marked tie rather than an improvement — see significance bands.
188 further models at this delivery mode carry a price but no independent score. They are absent from the comparison in both directions — unrated is not a zero, and an unmeasured model is neither an alternative nor a worse buy.
Retired models are never offered as an alternative, and a model only competes against its own delivery mode: batch trades latency for price, so it is not a like-for-like swap.
Nothing here measures latency, throughput, rate limits or how a model behaves on your prompts. Two models with the same index score are not interchangeable.
The models that nothing beats on both axes are on the value frontier, and every other model something cheaper beats is listed here.