Inference
LMArena

Agent Arena task outcome

Estimated effect of the model on whether a user marks an agent session as finished.

ModelsModels with a result.43
Top resultBest result on this test.19.8Claude Fable 5.1
Top-3 spreadPoints from first to third.6.5
UpdatedDate the source changed the results.15 Sept 2026

Result and price

-30-20-1001020$0.3$1$3$10$30$100RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

43 results
#ModelResultThe score from the source.
01Claude Fable 5.1max effort19.8
02GPT-6 Astramax effort17.7
03DeepSeek V4.1 Flashmax effort13.4
04Claude Opus 5max effort12.4
05Muse Spark 1.3max effort10.3
06Hy4 preview9.8
07Gemini 3.8 Flashhigh effort9.3
08Qwen3.8-Flash-Next8.9
09GLM-5.3max effort8.5
10GLM-5.3-Flash8.2
11DeepSeek V4 Pro (High) (0813)6.8
12Claude Opus 4.8high effort6.3
13Claude Fable 5high effort6.0
14Qwen3.8max effort5.8
15GLM-5.2max effort4.9
16Qwen3.8-27B4.3
17GPT-5.6 Solxhigh effort3.7
18Deepseek V4 Flash (High) (20260731)3.2
19Gemini 3.7 Flashhigh effort2.9
20Claude Sonnet 5high effort2.9
21Grok 4.51.9
22Muse Spark 1.10.0
23GPT-5.5xhigh effort-0.6
24GPT-5.4high effort-1.6
25DeepSeek V4 Pro 0423-1.7
26Muse Spark 1.2xhigh effort-2.0
27Grok 4.6xhigh effort-2.5
28GPT-5.6 Terraxhigh effort-4.4
29GPT-5.6 Lunaxhigh effort-5.0
30Gemini 3.1 Pro Preview-5.5
31Claude Sonnet 4.6-5.8
32Qwen3.7max effort-6.4
33Gemini 3.6 Flashhigh effort-6.8
34Qwen3.7 Plus-7.1
35MiMo-V2.5-Pro-9.0
36Hy3-10.6
37MiniMax M3-12.0
38Mistral Medium 3.5-15.1
39Gemini 3.5 Flash-Lite-17.6
40MiniMax M2.7-18.0
41Inkling-19.3
42Solar Pro 4-20.1
43Inkling-Small-20.4