Telegram

Leaderboards

Which model is best, for what.

One ranking per job, not one ranking for everything. Rebuilt every day from data the benchmark operators publish themselves — each board names its source and the date that source last updated.

Last rebuilt 2026-09-02 22:45 UTC

Best overall chat model

Head-to-head human preference across everyday prompts. The closest thing there is to a general ranking.

LMArena 2026-09-01

Rank# Model Made by Arena score Votes
1 Claude Opus 5 Max Anthropic 1,505 16,839
2 Claude Opus 5 High Anthropic 1,504 34,617
3 Claude Opus 4.6 High Anthropic 1,503 72,110
4 Claude Opus 4.6 Anthropic 1,498 76,032
5 Claude Fable 5 Anthropic 1,494 26,977
6 Gemini 3.7 Flash High Google 1,491 5,685
7 Claude Opus 4.7 High Anthropic 1,490 60,123
8 muse-spark-1.2 (xHigh) Meta 1,488 3,244
9 Gemini 3.5 Flash High Google 1,484 33,874
10 Claude Opus 4.7 Anthropic 1,483 61,256
11 Gemini 3.1 Pro Preview Google 1,480 102,763
12 Qwen3.8 Max Alibaba 1,480 13,136
13 Gemini 3 Pro Google 1,480 40,675
14 Muse Spark 1.1 Meta 1,479 23,768
15 Gemini 3.6 Flash High Google 1,476 22,009
16 Kimi K3 Max Open Moonshot AI 1,476 17,895
17 Gemini 3.5 Flash Medium Google 1,476 32,274
18 Glm 5.3 Max Open Zai 1,474 7,401
19 Qwen3.7 Max Preview Alibaba 1,474 3,710
20 Muse Spark Meta 1,474 13,571
21 GPT 5.5 High OpenAI 1,472 63,265
22 Qwen3.5 Max Preview Alibaba 1,471 21,487
23 Glm 5.3 Flash Open Zai 1,471 4,399
24 GPT 5.4 High OpenAI 1,470 60,657
25 Ernie 5.1 Baidu 1,468 37,137

Best model for coding

Judged on real build tasks where a human picks the better result, not on a self-reported pass rate.

LMArena 2026-09-01

Rank# Model Made by Arena score Votes
1 qwen3.8-max-0902 Alibaba 1,691 1,389
2 Claude Opus 5 Max Anthropic 1,688 10,334
3 Kimi K3 Max Open Moonshot AI 1,674 4,544
4 qwen3.8-max Alibaba 1,669 3,219
5 Claude Opus 5 High Anthropic 1,661 10,326
6 Hy4 Preview Open Tencent 1,629 1,368
7 Grok 4.6 High xAI 1,629 1,532
8 Claude Fable 5 Anthropic 1,628 9,066
9 Qwen3.8 Flash Next Open Alibaba 1,620 1,876
10 gpt-5.6-sol-xhigh (codex-harness) OpenAI 1,616 10,684
11 Glm 5.3 Max Open Zai 1,608 2,648
12 Glm 5.3 Flash Open Zai 1,604 1,698
13 Qwen3.8 27b Open Alibaba 1,599 3,256
14 Gemini 3.7 Flash High Google 1,587 3,007
15 Glm 5.2 Max Open Zai 1,585 9,569
16 Deepseek v4 Pro High Open DeepSeek 1,584 3,289
17 Deepseek v4 Flash High Open DeepSeek 1,581 3,928
18 Claude Opus 4.8 High Anthropic 1,563 12,603
19 Claude Opus 4.7 Anthropic 1,557 15,430
20 Claude Opus 4.7 High Anthropic 1,556 15,909
21 Grok 4.5 xAI 1,555 7,126
22 Claude Opus 4.6 High Anthropic 1,546 17,908
23 Claude Opus 4.8 Anthropic 1,540 11,543
24 Muse Spark 1.1 Meta 1,540 7,091
25 Gemini 3.6 Flash High Google 1,538 7,726

Best agent model

Long-running tasks with tool calls, where being nearly right at step three ruins step nine.

LMArena 2026-08-31

Rank# Model Made by Score % Votes
1 Claude Opus 5 (High) Anthropic 13.8 21,433
2 Claude Opus 5 (Max) Anthropic 11.6 17,073
3 Claude Fable 5 (High) Anthropic 10.6 34,905
4 GPT 5.6 Sol (xHigh) OpenAI 9.8 28,524
5 Claude Opus 4.8 (High) Anthropic 9.5 36,887
6 Kimi K3 (Max) Open Moonshot AI 8.7 94,549
7 GPT 5.5 (xHigh) OpenAI 7.9 50,040
8 Claude Sonnet 5 (High) Anthropic 7.5 27,467
9 Claude Opus 4.7 (High) Anthropic 6.6 36,570
10 Claude Opus 4.7 Anthropic 6.3 37,121
11 GPT 5.5 (High) OpenAI 6.1 73,880
12 GLM 5.2 (Max) Open Zai 6.1 65,079
13 Grok 4.5 xAI 6.1 34,008
14 Qwen3.8 Max Alibaba 6.0 18,386
15 DeepSeek V4 Pro (High) (0813) Open DeepSeek 5.9 23,928
16 Grok 4.6 (xHigh) xAI 5.8 15,090
17 Claude Opus 4.6 Anthropic 5.2 36,365
18 GPT 5.5 OpenAI 4.8 77,762
19 GLM 5.3 Flash Open Zai 4.4 9,970
20 GLM 5.3 (Max) Open Zai 3.8 41,022
21 GPT 5.4 (High) OpenAI 3.2 76,973
22 Deepseek V4 Flash (High) (20260731) Open DeepSeek 3.0 51,219
23 GPT 5.6 Terra (xHigh) OpenAI 2.9 16,778
24 Qwen3.8 Flash Next Open Alibaba 2.4 8,777
25 Claude Opus 4.8 Anthropic 2.3 33,690

Best model for images and documents

Understanding what is in a picture, a chart or a scanned page.

LMArena 2026-08-27

Rank# Model Made by Arena score Votes
1 Claude Fable 5 Anthropic 1,329 10,002
2 Claude Opus 5 High Anthropic 1,322 8,271
3 Claude Opus 4.7 Anthropic 1,317 21,457
4 Claude Opus 4.7 High Anthropic 1,316 21,137
5 Claude Opus 4.6 High Anthropic 1,315 20,883
6 Qwen3.8 Max Alibaba 1,313 7,244
7 Claude Opus 4.6 Anthropic 1,311 25,210
8 Gemini 3.5 Flash High Google 1,311 8,259
9 Muse Spark Meta 1,306 5,585
10 Gemini 3.5 Flash Medium Google 1,306 8,261
11 Gemini 3 Pro Google 1,304 13,019
12 muse-spark-1.2 (xHigh) Meta 1,304 1,844
13 Gemini 3.6 Flash High Google 1,302 3,816
14 GPT 5.4 High OpenAI 1,298 23,263
15 Glm 5.3 Flash Open Zai 1,296 1,389
16 GPT 5.5 OpenAI 1,296 21,389
17 Gemini 3.1 Pro Preview Google 1,295 39,412
18 GPT 5.5 High OpenAI 1,294 20,121
19 Claude Opus 4.8 High Anthropic 1,294 13,521
20 GPT 5.4 OpenAI 1,293 21,245
21 Muse Spark 1.1 Meta 1,293 6,769
22 Grok 4.5 xAI 1,291 6,401
23 Claude Opus 4.8 Anthropic 1,289 13,980
24 Gemini 3 Flash Google 1,285 36,747
25 Claude Sonnet 4.6 Anthropic 1,282 25,610

Best image generator

Text-to-image, ranked by people choosing between two results.

LMArena 2026-08-07

Rank# Model Made by Arena score Votes
1 gpt-image-2 (medium) OpenAI 1,380 69,194
2 Reve 2.1 Reve 1,302 7,598
3 Muse Image Meta 1,283 14,510
4 Reve 2.0 Reve 1,270 14,650
5 gemini-3.1-flash-image (nano-banana-2) [web-search] Google 1,263 27,910
6 Qwen Image 3.0 Pro Alibaba 1,258 3,735
7 Seedream 5.0 Pro ByteDance 1,257 26,510
8 Mai Image 2.5 Microsoft-Ai 1,256 44,940
9 gemini-3.1-flash-lite-image (nano-banana-2-lite) Google 1,251 16,419
10 gemini-3-pro-image-2k (nano-banana-pro) Google 1,246 138,756
11 GPT Image 1.5 High Fidelity OpenAI 1,239 142,427
12 gemini-3-pro-image-preview (nano-banana-pro) Google 1,232 82,835
13 Grok Imagine Image Quality xAI 1,228 48,786
14 Ideogram 4.0 Quality Open Ideogram 1,206 27,891
15 Qwen Image 2.0 Pro 2026.06.22 Alibaba 1,191 12,026
16 Uni 1.1 Max Luma AI 1,188 13,491
17 Mai Image 2 Microsoft-Ai 1,182 49,207
18 Cosmos3-Super-Text2Image (Agentic) Open NVIDIA 1,181 2,957
19 Uni 1.1 Luma AI 1,180 26,927
20 Grok Imagine Image xAI 1,172 217,461
21 Recraft V4.1 Utility Pro Recraft 1,169 2,519
22 Flux 2 Max Bfl 1,162 117,464
23 Grok Imagine Image Pro xAI 1,161 93,711
24 Cosmos3-Super-Text2Image Open NVIDIA 1,158 4,669
25 Flux 2 Flex Bfl 1,156 149,402

Best video generator

Text-to-video, same method: two clips, a human picks one.

LMArena 2026-08-27

Rank# Model Made by Arena score Votes
1 Gemini Omni 1.1 Flash Google 1,515 1,762
2 Gemini Omni Flash Google 1,512 19,830
3 Flux 3 Video Bfl 1,495 1,287
4 Dreamina Seedance 2.0 720p ByteDance 1,479 51,266
5 Dreamina Seedance 2.5 720p ByteDance 1,476 2,121
6 Minimax H3 Open MiniMax 1,460 5,813
7 Muse Video Meta 1,457 2,178
8 Happyhorse 1.0 Aorizon 1,428 22,112
9 Sora 2 Pro OpenAI 1,365 48,782
10 Veo 3.1 Audio Google 1,364 13,705
11 Veo 3.1 Audio 1080p Google 1,363 24,972
12 Veo 3.1 Fast Audio Google 1,361 39,371
13 Veo 3.1 Fast Audio 1080p Google 1,358 25,768
14 Veo 3 Fast Audio Google 1,347 25,169
15 Wan2.7 T2v Wan 1,344 20,882
16 Grok Imagine Video 720p xAI 1,344 157,732
17 Sora 2 OpenAI 1,341 65,433
18 Veo 3 Audio Google 1,340 18,941
19 Wan2.6 T2v Alibaba 1,329 47,267
20 Seedance V1.5 Pro ByteDance 1,256 75,890
21 Veo 3 Google 1,253 14,951
22 Veo 3 Fast Google 1,248 15,225
23 Wan2.5 T2v Preview Alibaba 1,247 30,497
24 Pixverse V5.6 Unknown 1,240 32,296
25 Runway Gen 4.5 Runway 1,224 41,583

Cheapest capable models

Live list prices across providers. Ranking here is arithmetic, not opinion.

OpenRouter 2026-09-02

Rank# Model Made by Per 1M tokens Context
1 Mistral Nemo Mistral $0.022 131K
2 Ling-3.0-flash — $0.032 262K
3 Granite 4.0 Micro IBM $0.041 131K
4 Nex-N2-Mini Nex AGI $0.044 262K
5 Solar Pro 4 Upstage $0.052 524K
6 Qwen3.7 Flash Qwen $0.055 1M
7 gpt-oss-20b OpenAI $0.055 131K
8 Llama 3.1 8B Instruct Meta $0.057 131K
9 Nova Micro 1.0 Amazon $0.061 128K
10 Granite 4.1 8B IBM $0.062 131K
11 Gemma 3 4B Google $0.062 131K
12 Command R7B (12-2024) Cohere $0.066 128K
13 Mercury 2.5 Preview Inception $0.068 260K
14 GPT-5 Nano (batch) OpenAI $0.069 400K
15 gpt-oss-120b OpenAI $0.070 131K
16 Laguna XS 2.1 Poolside $0.075 262K
17 Gemma 3 12B Google $0.075 131K
18 DeepSeek V4 Flash Latest — $0.077 1.311M
19 Qwen3 30B A3B Instruct 2507 Qwen $0.084 262K
20 Nemotron 3 Nano 30B A3B NVIDIA $0.087 262K
21 gpt-oss-20b (batch) OpenAI $0.087 131K
22 Gemini 2.5 Flash Lite (batch) Google $0.087 1.049M
23 GPT-4.1 Nano (batch) OpenAI $0.087 1.048M
24 DeepSeek V4 Flash 0731 DeepSeek $0.094 1.311M
25 DeepSeek V4 Flash 0423 DeepSeek $0.099 1.049M

How to read these

Arena scores are relative, not absolute. They come from people comparing two anonymous answers and picking one. A 20-point gap near the top is often inside the margin of error — treat the top few as a tie unless the confidence intervals separate.

A benchmark is not your workload. The board tells you which model wins on average across other people's prompts. Yours are different. Use it to pick two candidates, then test both on your own work.

Prices move faster than rankings. The price board is list price per million tokens, blended three parts input to one part output. Providers discount, cache and rate-limit differently, so it is a starting point, not a bill.

We do not run these tests. Every number here is published by the operator named next to it and copied without adjustment. If a board looks stale, its source has not updated — the date shown is theirs, not ours.