Every frontier model, ranked live

The Godot Model Index is a weighted composite of LMSYS Arena Elo (human preference), OpenRouter adoption rank, and pricing efficiency.

UPDATED · Sat, 17 Oct 2026 22:21:25 UTC

Scoring methodology →

#	Model	Godot	Elo	Weekly	$/M in	$/M out	Context	Category
1	Anthropic: Claude Opus 4.6 Anthropic	99.3	1498	#19	5.00	25.00	1000K	Multimodal
2	Z.ai: GLM 5.1 Z-ai	95.3	1462	#27	0.966	3.04	205K	Frontier
3	Qwen: Qwen3.5 Plus 2026-02-15 Alibaba	93.7	1470	#135	0.260	1.56	1000K	Multimodal
4	Google: Gemini 3 Flash Preview Google DeepMind	93.6	1442	#11	0.500	3.00	1049K	Multimodal
5	Anthropic: Claude Sonnet 4.6 Anthropic	93.4	1458	#7	3.00	15.00	1000K	Multimodal
6	Z.ai: GLM 5 Z-ai	92.6	1446	#40	1.00	3.20	205K	Frontier
7	SpaceXAI: Grok 4.20 xAI	91.0	1445	#97	1.25	2.50	2000K	Multimodal
8	Qwen: Qwen3.5 397B A17B Alibaba	90.8	1438	#57	0.500	3.60	262K	Multimodal
9	Anthropic: Claude Opus 4.5 Anthropic	90.7	1451	#75	5.00	25.00	200K	Multimodal
10	Z.ai: GLM 4.7 Z-ai	90.3	1435	#54	0.540	1.98	205K	Frontier
11	DeepSeek: DeepSeek V3.2 DeepSeek	90.2	1424	#10	0.280	0.420	164K	Open Source
12	Anthropic: Claude Sonnet 4.5 Anthropic	90.1	1438	#46	3.00	15.00	1000K	Multimodal
13	Google: Gemini 3.1 Flash Lite Preview Google DeepMind	89.7	1416	#37	0.250	1.50	1049K	Multimodal
14	MoonshotAI: Kimi K2.5 Moonshotai	89.7	1420	#44	0.500	2.50	262K	Multimodal
15	Z.ai: GLM 4.6 Z-ai	89.0	1440	#144	0.500	2.00	205K	Frontier
16	OpenAI: GPT-5.4 Mini OpenAI	88.8	1411	#36	0.750	4.50	400K	Multimodal
17	xAI: Grok 4.1 Fast xAI	88.7	1408	#455	0.200	0.500	2000K	Multimodal
18	Qwen: Qwen3 235B A22B Instruct 2507 Alibaba	88.4	1419	#28	0.090	0.550	262K	Open Source
19	Google: Gemini 2.5 Flash Google DeepMind	88.3	1406	#13	0.300	2.50	1049K	Multimodal
20	Qwen: Qwen3 VL 235B A22B Instruct Alibaba	87.7	1420	#112	0.210	1.90	262K	Multimodal
21	DeepSeek: DeepSeek V3.1 DeepSeek	87.6	1417	#66	0.250	0.950	164K	Open Source
22	OpenAI: GPT-5.1 OpenAI	87.3	1422	#47	1.25	10.00	400K	Multimodal
23	Xiaomi: MiMo-V2-Pro Xiaomi	87.3	1395	#436	1.00	3.00	1049K	Frontier
24	Google: Gemma 4 26B A4B Google DeepMind	87.3	1434	#155	0.068	0.225	262K	Multimodal
25	SpaceXAI: Grok 4.20 Multi-Agent xAI	87.2	1450	#182	1.25	2.50	2000K	Multimodal
26	DeepSeek: DeepSeek V3.2 Exp DeepSeek	86.8	1422	#100	0.270	0.410	164K	Open Source
27	OpenAI: GPT-5.2 OpenAI	86.8	1412	#71	1.75	14.00	400K	Multimodal
28	MiniMax: MiniMax M2.7 Minimax	86.6	1392	#26	0.300	1.20	205K	Frontier
29	DeepSeek: DeepSeek V3.1 Terminus DeepSeek	86.5	1417	#92	0.270	1.00	164K	Open Source
30	DeepSeek: R1 0528 DeepSeek	86.5	1428	#109	0.500	2.15	164K	Reasoning
31	Qwen: Qwen3.5-Flash Alibaba	86.4	1397	#59	0.065	0.260	1000K	Multimodal
32	OpenAI: GPT-5.2 Chat OpenAI	86.3	1438	#171	1.75	14.00	128K	Multimodal
33	Qwen: Qwen3.5-122B-A10B Alibaba	86.3	1417	#141	0.260	2.08	262K	Multimodal
34	xAI: Grok 4 Fast xAI	86.2	1400	#468	0.200	0.500	2000K	Multimodal
35	OpenAI: GPT-5 Chat OpenAI	86.1	1403	#200	1.25	10.00	128K	Multimodal
36	Anthropic: Claude Haiku 4.5 Anthropic	85.4	1397	#33	1.00	5.00	200K	Multimodal
37	OpenAI: GPT-5.4 Pro OpenAI	84.9	1469	#295	30.00	180.00	1050K	Multimodal
38	Qwen: Qwen3.5-27B Alibaba	84.8	1409	#83	0.260	2.60	262K	Multimodal
39	Qwen: Qwen3.5-35B-A3B Alibaba	84.6	1395	#68	0.150	1.00	262K	Multimodal
40	MoonshotAI: Kimi K2 Thinking Moonshotai	84.2	1414	#189	0.600	2.50	262K	Reasoning
41	Anthropic: Claude Opus 4.7 Anthropic	84.2	1377	#8	5.00	25.00	1000K	Multimodal
42	OpenAI: GPT-5.4 Nano OpenAI	83.9	1372	#49	0.200	1.25	400K	Multimodal
43	Google: Gemini 2.5 Flash Lite Google DeepMind	83.9	1369	#14	0.100	0.400	1049K	Multimodal
44	Qwen: Qwen3 235B A22B Thinking 2507 Alibaba	83.7	1415	#124	0.450	3.50	128K	Reasoning
45	Z.ai: GLM 4.5 Z-ai	83.5	1430	#208	0.600	2.20	131K	Frontier
46	OpenAI: GPT-5 Mini OpenAI	83.5	1373	#43	0.250	2.00	400K	Multimodal
47	Google: Gemini 2.5 Flash Lite Preview 09-2025 Google DeepMind	83.4	1379	#87	0.100	0.400	1049K	Multimodal
48	OpenAI: GPT-4.1 OpenAI	83.0	1383	#70	2.00	8.00	1048K	Multimodal
49	Qwen: Qwen3 30B A3B Instruct 2507 Alibaba	82.6	1384	#107	0.100	0.300	262K	Open Source
50	Baidu: ERNIE 4.5 300B A47B Baidu	82.1	1445	#428	0.280	1.10	131K	Frontier

Showing 50 of 50 models

Open-source pulse · HuggingFace

Trending right now

TOP 10 · UPDATED DAILY

01
embeddinggemma-2
google·feature-extraction·11d ago
151.3k
downloads
2.5k
likes
02
humanizer
jialinyyzz·text-generation·2d ago
85.6k
downloads
1.8k
likes
03
MiniCPM-V-4.7-35B-A3B
openbmb·image-text-to-text·2d agoNEW
17.6k
downloads
852
▲ 1
04
drama-3
fishaudio·text-to-speech·3d ago
6.2k
downloads
561
▲ 2
05
Qwen-Image-2.1-Turbo
Qwen·text-to-image·5d agoNEW
17.4k
downloads
946
▼ 2
06
LTX-2.5
Lightricks·image-to-video·6d ago
1.9M
downloads
8.1k
▲ 1
07
Qwen-Image-2.1-Uncensored-Quantized
abenzerps·text-to-image·5d ago
2.5M
downloads
4.4k
▼ 2
08
Muse-Spark-Open-816B-A42B
meta-models·image-text-to-text·2d ago
815
downloads
451
▲ 6
09
LFM2.5-24B-A2B
LiquidAI·text-generation·4d agoNEW
9.4k
downloads
436
likes
10
flux-3-image-klein-comfy
black-forest-labs·image-to-image·20h agoNEW
73.8k
downloads
464
▲ 8

Methodology

How the Godot Model Index is computed

Weighted composite on a 0–100 scale: 60% capability (LMSYS Arena Elo, min–max normalized across ranked models), 30% adoption (OpenRouter weekly usage rank across 680+ models), and 10% efficiency (inverse of per-token price). Models without an Arena signal are capped at 50. Updated daily via cron at 03:00 UTC.

· Capability signal: https://lmarena.ai/ (via HuggingFace dataset snapshot).
· Adoption signal: https://openrouter.ai/ frontend weekly ranking.
· Efficiency signal: OpenRouter pricing (input + output $/1M tokens).
· Name matching across sources uses a Haiku-backed alias table.