Chinese AI Model Rankings

Where the Chinese AI models ChinaAPI serves stand next to the international frontier, quoted from public leaderboards. Snapshot of 2026-08-26; roster last maintained 2026-08-26 (moves when a model or source is added). Every figure is reproduced from the board named in its column and links back to it. ChinaAPI runs no evaluation of its own and computes no composite score from these numbers. See how this list is built, the price of every model marked available, and the same data as JSON at /api/model-rankings.

Sources

Text and multimodal models

Ordered by Epoch Capabilities Index (ECI); a model with no figure on that board follows the ones that have one, because absent from a board is not the same as last on it.

Text and multimodal models, public benchmark results as of 2026-08-26
#ModelVendorOriginReleasedOn ChinaAPILMArena Text ArenaLiveBenchArtificial Analysis Intelligence IndexSuperCLUE 智能指数Epoch Capabilities Index (ECI)Vals IndexVendor's own evaluation cardNotes
1Claude Fable 5AnthropicInternational2026-06-09no1508 ±583.062162 (160 - 166)66.04 ±1.03
2Claude Opus 5AnthropicInternational2026-07-24no1493 ±580.163162 (159 - 165)67.21 ±0.98
3GPT-5.6 SolOpenAIInternational2026-07-09no1482 ±581.061161 (159 - 165)63.71 ±1.06
4GPT-5.5OpenAIInternational2026-04-23no1482 ±480.2159 (156 - 162)57.41 ±1.15
5GPT-5.6 TerraOpenAIInternational2026-07-09no1465 ±577.957159 (156 - 162)59.59 ±1.33
6Gemini 3.7 FlashGoogleInternational2026-08-13no1490 ±878.856157 (155 - 160)59.31 ±1.06Released 2026-08-13. LMArena marks the row preliminary (5,718 votes as of 2026-08-26).
7Kimi K3Moonshot AIChina2026-07-16yes1489 ±679.26070.68157 (155 - 161)57.81 ±1.06
8Grok 4.6xAIInternational2026-08-12no1461 ±1078.061156 (154 - 158)59.17 ±1.21Released 2026-08-12. LMArena labels this vendor SpaceXAI and marks the row preliminary (3,473 votes as of 2026-08-26).
9Qwen3.8-MaxAlibabaChina2026-07-19yes1481 ±778.55871.48156 (155 - 159)51.84 ±1.29
10Gemini 3.1 ProGoogleInternational2026-02-19no1486 ±377.048155 (153 - 157)41.90 ±1.17
11DeepSeek V4 Flash 0731DeepSeekChina2026-07-31yes74.25265.60154 (152 - 156)53.57 ±1.21The build the deepseek-v4-flash API serves.
12Gemini 3.6 FlashGoogleInternational2026-07-21yes1481 ±573.652154 (153 - 156)55.35 ±1.09
13Grok 4.5xAIInternational2026-07-08no1470 ±575.856154 (152 - 156)51.53 ±1.32LMArena labels this vendor SpaceXAI.
14Qwen3.7-MaxAlibabaChina2026-05-19yes1474 ±1073.1154 (152 - 156)44.77 ±1.08LMArena lists the preview build (qwen3.7-max-preview); the released model has no separate row.
15GLM-5.2Zhipu AI (Z.ai)China2026-06-16yes1470 ±573.25363.27152 (150 - 154)53.12 ±1.18
16GLM-5.1Zhipu AI (Z.ai)China2026-04-07yes1468 ±4151 (149 - 152)The Vals Index V2 (2026-08-13) no longer lists this model; the V1.2 figure it once carried was removed with that revision.
17Kimi K2.6Moonshot AIChina2026-04-20yes1461 ±570.5151 (149 - 153)43.47 ±1.17
18Kimi K2.7 CodeMoonshot AIChina2026-06-12yes68.443151 (149 - 152)
19DeepSeek V4 ProDeepSeekChina2026-04-24no1458 ±445149 (148 - 151)42.89 ±1.19The pre-0813 build; the deepseek-v4-pro API moved to DeepSeek-V4-Pro-0813 on 2026-08-12. LiveBench dropped this row in favour of the dated builds between the 2026-08-19 and 2026-08-26 snapshots, so its 71.6 is no longer quoted.
20Qwen3.6-PlusAlibabaChina2026-04-01yes68.9149 (147 - 151)31.98 ±0.98
21MiniMax-M3MiniMaxChina2026-06-01yes67.34556.90148 (143 - 150)42.72 ±1.20
22Qwen3.7-PlusAlibabaChina2026-06-02yes1457 ±539148 (146 - 151)38.65 ±1.17
23GLM-5Zhipu AI (Z.ai)China2026-02-11yes1457 ±4146 (144 - 148)
24MiniMax-M2.7MiniMaxChina2026-03-18yes146 (139 - 148)24.56 ±0.93
25Qwen3.6-FlashAlibabaChina2026-04-26yes144 (141 - 146)
26DeepSeek V4 Pro 0813DeepSeekChina2026-08-13yes1459 ±1077.45370.1052.37 ±1.14The build the deepseek-v4-pro API has served since 2026-08-12. LMArena's 0813 row now carries a rank and 3,621 human votes rather than the unranked AutoEval preview it was on 2026-08-18, so it is quoted here; Epoch still scores only the undated DeepSeek-V4-Pro and that figure is not reused for this build.
27MiMo-V2.5-ProXiaomiChina2026-04-22yes1468 ±44353.0140.97 ±1.38
28Muse Spark 1.2MetaInternational2026-08-05no1498 ±1078.05757.05 ±1.12
29GLM-5.3Zhipu AI (Z.ai)China2026-08-14yes1487 ±1076.160Released 2026-08-14; served here via Zhipu's official channels since 2026-08-18. Independent boards picked it up between the 2026-08-19 and 2026-08-26 snapshots, so the vendor's own DeepSWE figure it carried before has been replaced by their readings; the LMArena row stands on 3,751 votes. Epoch does not score it, which is why it sits after the models this category's primary board covers.
30Hunyuan Hy3TencentChina2026-07-06yes1457 ±84262.13
31DeepSeek V4 Flash Vision ExpDeepSeekChina2026-08-21yes76.851DeepSeek's experimental vision tier, announced in the vendor's change log on 2026-08-21 as model='deepseek-v4-flash-vision-exp'. LiveBench names its row with the vendor's -Exp suffix and Artificial Analysis drops it, but DeepSeek publishes only one vision tier, so both rows are this model. LMArena, Vals, SuperCLUE and Epoch carry no row for it, and the sibling flash figures are not reused; with no Epoch figure it sits after the models this category's primary board covers.
32LongCat-2.0MeituanChina2026-06-30yes3454.22
33MiMo-V2.5XiaomiChina2026-04-22yes3839.91 ±1.31
34Step 3.7 FlashStepFunChina2026-05-29yes3148.16
35DeepSeek V4 FlashDeepSeekChinano1436 ±4The pre-0731 build. LMArena still lists it alongside 0731; LiveBench dropped its row between the 2026-08-19 and 2026-08-26 snapshots, so its 65.5 is no longer quoted.
36Doubao-Seed-2.1-proByteDanceChina2026-06-28yes65.94Terminal Bench 2.1 71.0
37GLM-5V-TurboZhipu AI (Z.ai)Chinayes1433 ±7Vision-language model. LMArena's text board now carries a glm-5v-turbo row (9,380 votes); Epoch does not score it, so it sits after the models the primary board covers.
38Agnes 2.0 FlashAgnes AIChinayesClaw-Eval Pass^3 60.9%No independent board in this snapshot covers it; the figure shown is the vendor's own.
39Agnes 2.5 FlashAgnes AIChinayesArtificial Analysis carries a different build of this family (Agnes 2.5 Pro Alpha, 40) and the vendor publishes its own figures only as a chart image, so there is no number here that can be quoted.
40Doubao-Seed-2.1-turboByteDanceChina2026-06-28yesTerminal Bench 2.1 67.6No independent board in this snapshot covers it; the figure shown is the vendor's own. The vendor writes the name without the 260628 snapshot suffix we serve it under.
41GLM-5-TurboZhipu AI (Z.ai)China2026-03-16yesNo independent board in this snapshot carries a row for it.
42Step 3.5 FlashStepFunChina2026-02-02yesSWE-bench Verified 74.4No independent board in this snapshot covers it; the figure shown is the vendor's own.

Image generation models

Ordered by LMArena Text-to-Image Arena; a model with no figure on that board follows the ones that have one, because absent from a board is not the same as last on it.

Image generation models, public benchmark results as of 2026-08-26
#ModelVendorOriginReleasedOn ChinaAPILMArena Text-to-Image ArenaLMArena Image Edit ArenaNotes
1GPT Image 2OpenAIInternational2026-04-21no1382 ±41462 ±4
2Gemini 3.1 Flash ImageGoogleInternationalno1263 ±51387 ±4
3Doubao Seedream 5.0ByteDanceChina2026-01-28yes1258 ±51395 ±4Scores are the arena's seedream-5.0-pro rows; the arena also carries a seedream-5.0-lite build (1136 / 1293).
4FLUX.2 ProBlack Forest LabsInternational2025-11-25no1155 ±31245 ±3
5Doubao Seedream 4.5ByteDanceChina2025-11-28yes1147 ±31301 ±2
6Wan 2.7 Image ProAlibabaChina2026-04-01yes1103 ±51302 ±4
7Wan 2.7 ImageAlibabaChina2026-04-01yes1100 ±51301 ±4
8Agnes Image 2.0 FlashAgnes AIChinayesNo row on the LMArena image boards at the time of this snapshot.
9Agnes Image 2.1 FlashAgnes AIChinayesNo row on the LMArena image boards at the time of this snapshot.
10MiniMax image-01MiniMaxChinayesNo row on the LMArena image boards at the time of this snapshot.
11Step 2X LargeStepFunChinayesNo row on the LMArena image boards at the time of this snapshot; the arena carries an older StepFun build (step1x-edit, 998) that is a different model.
12Step Image Edit 2StepFunChinayesNo row on the LMArena image boards at the time of this snapshot; the arena carries an older StepFun build (step1x-edit, 998) that is a different model.

Video generation models

Ordered by LMArena Text-to-Video Arena; a model with no figure on that board follows the ones that have one, because absent from a board is not the same as last on it.

Video generation models, public benchmark results as of 2026-08-26
#ModelVendorOriginReleasedOn ChinaAPILMArena Text-to-Video ArenaLMArena Image-to-Video ArenaNotes
1Gemini Omni Flash (video)GoogleInternationalno1512 ±111462 ±6
2Doubao Seedance 2.0ByteDanceChina2026-01-28yes1482 ±101478 ±8The arena entry is the 720p Dreamina build of Seedance 2.0.
3Doubao Seedance 2.5ByteDanceChina2026-06-28yes1477 ±191484 ±12The LMArena video boards picked up dreamina-seedance-2.5-720p between the 2026-08-19 and 2026-08-26 snapshots. Both rows are still thin — 1,337 text-to-video and 2,912 image-to-video votes — and their intervals are correspondingly wide.
4MiniMax H3MiniMaxChina2026-07-31yes1453 ±131489 ±7
5Sora 2 ProOpenAIInternational2025-09-30no1364 ±7
6Veo 3.1GoogleInternational2025-10-15no1364 ±141397 ±11
7Wan 2.7 T2VAlibabaChina2026-04-03yes1344 ±9
8MiniMax Hailuo 2.3MiniMaxChina2025-10-28yes1204 ±71261 ±5
9MiniMax Hailuo 02MiniMaxChinayes1198 ±121228 ±11Scores are the arena's hailuo-02-pro rows; the arena also carries standard and fast builds of the same family.
10Kling v3KuaishouChina2026-02-05yes1356 ±6The arena entry is the pro build; the text-to-video board has no v3 row at the time of this snapshot.
11Wan 2.7 I2VAlibabaChina2026-04-03yes1427 ±5
12HappyHorse 1.1 I2VAlibabaChinayesThe LMArena video boards carry the previous generation (happyhorse-1.0, 1428 text-to-video / 1442 image-to-video); that score is not this model's and is deliberately not reused here.
13HappyHorse 1.1 T2VAlibabaChinayesThe LMArena video boards carry the previous generation (happyhorse-1.0, 1428 text-to-video / 1442 image-to-video); that score is not this model's and is deliberately not reused here.
14Kling 3.0 TurboKuaishouChinayesNo row on the LMArena video boards at the time of this snapshot.
15Kling v3 OmniKuaishouChina2026-02-05yesNo row on the LMArena video boards at the time of this snapshot.
16MiniMax Hailuo 2.3 FastMiniMaxChina2025-10-28yesNo row on the LMArena video boards at the time of this snapshot.

Real usage on ChinaAPI

Which models the calls ChinaAPI actually served went to, as positions and shares of this platform's own traffic. It is a measurement of one gateway, not of the Chinese AI market, and it is not comparable with the benchmark results above. No call count, token count or revenue figure is published here or anywhere else on this site: every number below is a proportion. Recomputed hourly. A model with too little traffic in a window to rank is left unnamed and counted under “other”, as is anything outside the published catalogue. The same board as JSON is at /api/model-rankings/usage.

Share of calls served, the last 7 days

Model share of calls ChinaAPI served over the last 7 days
#ModelVendorShare of callsTrend vs the previous 7 days
1deepseek-v4-flashDeepSeek10.4%down
2agnes-2.5-flashAgnes AI6.9%down
3deepseek-v4-proDeepSeek4.4%down
4gpt-5.6-solOpenAI4.2%flat
5gemini-3.6-flashGoogle3.0%down
6kimi-k2.7-codeMoonshot2.9%up
7kimi-k3Moonshot2.7%up
8gpt-5.6-lunaOpenAI2.3%down
9claude-opus-4-8Anthropic2.1%down
10glm-5.3-flash智谱2.0%up
11mimo-v2.5Xiaomi1.8%up
12deepseek-v4-flash-vision-expDeepSeek1.7%up
13gpt-5.6-terraOpenAI1.7%down
14agnes-image-2.1-flashAgnes AI1.7%up
15gpt-5.5OpenAI1.6%down
16claude-sonnet-5Anthropic1.5%down
17claude-opus-5Anthropic1.5%down
18glm-5.2智谱1.3%flat
19MiniMax-M2.7MiniMax1.3%flat
20claude-opus-4-7Anthropic1.3%flat
All other models43.6%

How those calls arrived: 9.2% over agent protocols (Claude Messages and the OpenAI Responses API), 86.2% over chat protocols (OpenAI chat completions and Gemini generateContent), 4.6% over everything else (audio, images, embeddings, rerank and asynchronous tasks).

Share of calls served, the last 30 days

Model share of calls ChinaAPI served over the last 30 days
#ModelVendorShare of callsTrend vs the previous 30 days
1deepseek-v4-flashDeepSeek21.6%down
2deepseek-v4-proDeepSeek14.5%down
3mimo-v2.5Xiaomi6.3%down
4agnes-2.5-flashAgnes AI3.9%up
5claude-sonnet-5Anthropic3.2%up
6gpt-5.6-solOpenAI2.9%up
7kimi-k3Moonshot2.5%down
8gemini-3.6-flashGoogle2.3%up
9gpt-5.6-lunaOpenAI2.3%up
10claude-opus-4-8Anthropic1.9%up
11gpt-5.6-terraOpenAI1.9%up
12claude-opus-5Anthropic1.7%up
13kimi-k2.7-codeMoonshot1.6%up
14gpt-5.5OpenAI1.6%up
15glm-5.2智谱1.5%down
16gemini-3.1-flash-liteGoogle1.2%up
17gpt-4.1-miniOpenAI1.1%up
18MiniMax-M2.7MiniMax0.9%down
19claude-fable-5Anthropic0.9%flat
20glm-5智谱0.9%flat
All other models25.2%

How those calls arrived: 6.3% over agent protocols (Claude Messages and the OpenAI Responses API), 67.3% over chat protocols (OpenAI chat completions and Gemini generateContent), 26.4% over everything else (audio, images, embeddings, rerank and asynchronous tasks).