Arena (formerly LMArena)

Name: Arena (formerly LMArena) results
Creator: Arena (formerly LMArena)
License: https://creativecommons.org/licenses/by/4.0/

Estimated effect on avoiding calls to tools that do not exist.

Last updated 2 Oct 2026

Results dated: 2 Oct 2026
Results: 50 configurations of 48 models
Unit: IPS effect estimate
Licence: Creative Commons Attribution 4.0 International

Tool grounding: Claude Fable 5.1

Top 15 of 50 results · IPS effect estimate, higher is better · lines show the 95% range · ≈ cannot be told apart from the leader. Choose a model to highlight it.Clear highlight

1≈ GPT-6.1 Sol (max reasoning)OpenAI 0.00
1≈ Kimi K3 (max reasoning)Moonshot AI 0.00
1≈ Claude Fable 5 (high reasoning)Anthropic 0.00
1≈ Claude Fable 5.1 (max reasoning)Anthropic 0.00
1≈ Claude Opus 5 (high reasoning)Anthropic 0.00
1≈ Claude Opus 5.5 (high reasoning)Anthropic 0.00
1≈ Claude Sonnet 5.5 (max reasoning)Anthropic 0.00
1≈ Step 5 PreviewStepFun 0.00
1≈ Claude Opus 5 (max reasoning)Anthropic 0.00
1≈ GLM-5.2 (max reasoning)Z.ai 0.00
1≈ Gemini 3.6 Flash (high reasoning)Google 0.00
1≈ Grok 4.5xAI 0.00
1≈ GLM-5.3 (max reasoning)Z.ai 0.00
1≈ GPT-6 Astra (max reasoning)OpenAI 0.00
1≈ Muse Spark 1.3 (max reasoning)Meta 0.00

Full results

Arena Agent: tool grounding, IPS effect estimate, higher is better
#	Model	Tool grounding · 95% range IPS effect estimate, higher is better	Price $ per million tokens, in / out
1≈	GPT-6.1 Sol (max reasoning)OpenAI	0.00 0.00–0.01	$2 / $10
1≈	Kimi K3 (max reasoning)Moonshot AI	0.00 0.00–0.01	$3 / $15
1≈	Claude Fable 5 (high reasoning)Anthropic	0.00 0.00–0.01	$10 / $50
1≈	Claude Fable 5.1 (max reasoning)Anthropic	0.00 0.00–0.01	$10 / $50
1≈	Claude Opus 5 (high reasoning)Anthropic	0.00 0.00–0.01	$5 / $25
1≈	Claude Opus 5.5 (high reasoning)Anthropic	0.00 0.00–0.01	$4 / $20
1≈	Claude Sonnet 5.5 (max reasoning)Anthropic	0.00 0.00–0.01	$2 / $10
1≈	Step 5 PreviewStepFun	0.00 0.00–0.01	–
1≈	Claude Opus 5 (max reasoning)Anthropic	0.00 0.00–0.01	$5 / $25
1≈	GLM-5.2 (max reasoning)Z.ai	0.00 0.00–0.00	$1.40 / $4.40
1≈	Gemini 3.6 Flash (high reasoning)Google	0.00 0.00–0.00	$1.50 / $7.50
1≈	Grok 4.5xAI	0.00 0.00–0.01	$2 / $6
1≈	GLM-5.3 (max reasoning)Z.ai	0.00 0.00–0.01	$1.40 / $4.40
1≈	GPT-6 Astra (max reasoning)OpenAI	0.00 0.00–0.01	$10 / $50
1≈	Muse Spark 1.3 (max reasoning)Meta	0.00 0.00–0.00	$1.25 / $4.25
1≈	GLM-5.3-FlashZ.ai	0.00 0.00–0.00	$0.15 / $0.50
1≈	DeepSeek-V4.1-Flash (max reasoning)DeepSeek	0.00 0.00–0.00	$0.30 / $1.20
1≈	Gemini 3.8 Flash (high reasoning)Google	0.00 0.00–0.00	$1.50 / $7.50
1≈	GPT-5.5 (extra-high reasoning)OpenAI	0.00 0.00–0.00	$5 / $30
1≈	Claude Sonnet 5 (high reasoning)Anthropic	0.00 0.00–0.00	$2 / $10
1≈	Grok 4.6 (extra-high reasoning)xAI	0.00 0.00–0.00	$2 / $6
1≈	DeepSeek-V4-Pro (0813, high reasoning)DeepSeek	0.00 0.00–0.00	$1.32 / $3.96
1≈	GPT-5.6 Luna (extra-high reasoning)OpenAI	0.00 0.00–0.00	$0.20 / $1.20
1≈	GPT-5.5OpenAI	0.00 0.00–0.00	$5 / $30
1≈	Grok 4.7 (extra-high reasoning)xAI	0.00 0.00–0.00	$2 / $6
1≈	Claude Opus 4.8 (high reasoning)Anthropic	0.00 0.00–0.00	$5 / $25
1≈	Gemini 3.7 Flash (high reasoning)Google	0.00 0.00–0.00	$1.50 / $7.50
1≈	GPT-5.6 Sol (extra-high reasoning)OpenAI	0.00 0.00–0.00	$4 / $20
1≈	InklingThinking Machines	0.00 0.00–0.00	$0.95 / $4.05
1≈	GPT-5.4 (high reasoning)OpenAI	0.00 -0.00–0.00	$2.50 / $15
1≈	Gemini 4 Argon (high reasoning)Google	0.00 -0.00–0.00	–
1≈	GPT-6 Sol (max reasoning)OpenAI	0.00 -0.00–0.00	$2 / $10
1≈	Hy4 PreviewTencent	0.00 -0.00–0.00	$0.83 / $2.50
1≈	MiniMax-M3MiniMax	0.00 -0.00–0.00	$0.30 / $1.20
1≈	GPT-5.6 Terra (extra-high reasoning)OpenAI	0.00 -0.00–0.00	$2 / $12
1≈	Qwen3.7-MaxAlibaba	0.00 -0.00–0.00	$1.48 / $4.42
1≈	Gemini 3.1 Pro PreviewGoogle	-0.00 -0.01–0.01	$2 / $12
1≈	Muse Spark 1.2 (extra-high reasoning)Meta	-0.00 -0.00–-0.00	$1.25 / $4.25
1≈	Muse Spark 1.1Meta	-0.00 -0.00–-0.00	$1.25 / $4.25
1≈	Qwen3.8-27BAlibaba	-0.00 -0.00–-0.00	$0.50 / $3
1≈	GPT-6 Luna (max reasoning)OpenAI	-0.00 -0.00–-0.00	$0.10 / $0.50
1≈	Inkling SmallThinking Machines	-0.00 -0.01–-0.00	$0.45 / $1.20
43	Qwen3.8-Flash-NextAlibaba	-0.01 -0.01–-0.01	–
43	Qwen3.7-PlusAlibaba	-0.01 -0.01–-0.00	$0.32 / $1.28
43	Qwen3.8-Max (0902)Alibaba	-0.01 -0.01–-0.01	$2 / $6
43	MiMo-V2.5-ProXiaomi	-0.01 -0.01–-0.01	$0.43 / $0.87
43	Solar Pro 4Upstage	-0.01 -0.02–-0.01	$0.09 / $0.36
48	Mistral Medium 3.5Mistral AI	-0.02 -0.03–-0.01	$1.50 / $7.50
49	Hy3Tencent	-0.03 -0.04–-0.02	$0.14 / $0.58
50	MiMo-V2.6-FlashXiaomi	-0.04 -0.04–-0.03	$0.14 / $0.28

Swipe the table sideways for more columns.

Ranks follow the score as shown, so equal numbers share a rank. ≈ marks results whose 95% range overlaps the leader's: they cannot be told apart from it. Results as published by Arena (formerly LMArena); we do not re-run them.

What it measures

Estimated effect on avoiding calls to tools that do not exist.

What it does not measure

Not a success rate: the estimated effect of choosing this model as the agent's orchestrator, against an even mix of the models tested, inside Arena's own harness.

Source

Arena (formerly LMArena) · Methodology

Contains data from the Arena Leaderboard Dataset by Arena, licensed under CC BY 4.0. Licence: Creative Commons Attribution 4.0 International.