IT之家 News on October 12: The Arena (arena.ai) platform has released its Week 41, 2026 AI large model rankings, covering 2026-10-05 ~ 2026-10-11. In this edition, Anthropic's brand-new Claude Opus 5.5 Max took first place in its debut appearance on the agent chart; the domestic model kimi-k3-max held steady in the top 20 of the overall chart, and multiple domestic models such as qwen3.8-max maintained stable rankings, sitting overall in the upper-middle tier.
Note from IT之家: This ranking is based on anonymous blind-test voting on the Arena (arena.ai) platform, with rankings calculated via ELO (the agent chart uses a percentage signal); it reflects users' subjective preferences rather than absolute capability testing; the different sub-charts are independent of each other and should not be compared across charts; score differences within the margin of error are treated as ties, and new models also need to accumulate a certain number of votes before officially entering the chart.
Overall Chart
The top of this week's overall chart was relatively stable, with gemini-4-argon-high continuing to hold first place and keeping an ELO score of 1525. Multiple Anthropic models occupy the top four spots: claude-opus-5.5-high rose 2 places to second with an ELO score of 1507 — still a gap from the top, but the difference is within 30 points, so it is considered close to third-place claude-opus-4-6-high within the margin of error. Two new models were added this edition, gpt-5.5 and gemini-3.5-flash-medium, ranking 29th and 30th respectively with ELO scores of 1477 and 1476.
Domestic models overall sit in the upper-middle range, with a relatively stable lineup:
Moonshot AI's kimi-k3-max ranks highest at 16th place with an ELO of 1488, unchanged;
Alibaba's qwen3.8-max is 22nd with an ELO of 1483, up 1 place from last week;
Xiaomi's mimo-v2.6-pro is 26th with an ELO of 1480, unchanged;
Zhipu's glm-5.3-max is 27th with an ELO of 1478, unchanged.
| Rank | vs. last week | Model | Vendor | Score (±CI) | Votes | Price ($/M) Input / Output | Context |
|---|---|---|---|---|---|---|---|
| 1 | - | gemini-4-argon-high | 1525 ±9 | 4892 | $2 / $10 | 1M | |
| 2 | ↑ 2 | claude-opus-5.5-high | Anthropic | 1507 ±8 | 6272 | $4 / $20 | 1M |
| 3 | ↓ 1 | claude-opus-4-6-high | Anthropic | 1504 ±3 | 79415 | $5 / $25 | 1M |
| 4 | ↓ 1 | claude-fable-5-high | Anthropic | 1504 ±4 | 40026 | $10 / $50 | 1M |
| 5 | - | claude-opus-4-7-high | Anthropic | 1501 ±4 | 65192 | $5 / $25 | 1M |
| 6 | - | claude-fable-5.1-max | Anthropic | 1501 ±6 | 13063 | $10 / $50 | 1M |
| 7 | - | claude-opus-4-6 | Anthropic | 1498 ±3 | 84095 | $5 / $25 | 1M |
| 8 | - | gemini-3.8-flash-high | 1497 ±5 | 31437 | $0.75 / $3.75 | 1.05M | |
| 9 | - | muse-spark-1.3-max | Meta | 1494 ±6 | 14694 | $1.25 / $4.25 | 1M |
| 10 | - | claude-opus-4-7 | Anthropic | 1494 ±4 | 66315 | $5 / $25 | 1M |
| — The following are domestic models outside the Top 10 — | |||||||
| 16 | - | kimi-k3-max | Moonshot | 1488 ±5 | 30351 | N/A | N/A |
| 22 | ↑ 1 | qwen3.8-max | Alibaba | 1483 ±5 | 25378 | $2 / $6 | 1M |
| 26 | - | mimo-v2.6-pro | Xiaomi | 1480 ±9 | 4069 | $0.44 / $0.87 | 1.05M |
| 27 | - | glm-5.3-max | Z.ai | 1478 ±5 | 20049 | $1.4 / $4.4 | 1.05M |
Code Leaderboard
On this week's code leaderboard, gemini-4-argon-high continues to hold the top spot, with its ELO score rising slightly by 2 points to 1562. Anthropic models showed significant ranking changes: claude-opus-5.5-high surged 7 places into the top five, reaching an ELO score of 1550, just 1 point behind second-place claude-fable-5-high (1551), which counts as a statistical tie. Zhipu's glm-5.3-flash rose 2 places to No. 27 with an ELO of 1525, becoming the highest-ranked domestic model. This edition adds the claude-opus-4-5-20251101 model, which debuts at No. 30 with an ELO of 1523.
The specific performance of domestic models on the code leaderboard is as follows:
Moonshot AI's kimi-k3-max ranks No. 10 with an ELO of 1540, down 1 place from last week;
Xiaomi's mimo-v2.6-pro ranks No. 11 with an ELO of 1539, down 1 place from last week;
DeepSeek's deepseek-v4.1-flash-max ranks No. 26 with an ELO of 1528, unchanged in rank;
Zhipu's glm-5.3-flash ranks No. 27 with an ELO of 1525, up 2 places from last week;
Alibaba's qwen3.8-max ranks No. 29 with an ELO of 1524, down 1 place from last week.
| Rank | vs last week | Model | Vendor | Score (±CI) | Votes | Price ($/M) Input / Output | Context |
|---|---|---|---|---|---|---|---|
| 1 | - | gemini-4-argon-high | 1562 ±17 | 1233 | $2 / $10 | 1M | |
| 2 | - | claude-fable-5-high | Anthropic | 1551 ±7 | 10404 | $10 / $50 | 1M |
| 3 | ↑ 1 | claude-opus-4-7-high | Anthropic | 1550 ±6 | 18591 | $5 / $25 | 1M |
| 4 | ↓ 1 | claude-opus-4-6-high | Anthropic | 1550 ±6 | 20617 | $5 / $25 | 1M |
| 5 | ↑ 7 | claude-opus-5.5-high | Anthropic | 1550 ±16 | 1444 | $4 / $20 | 1M |
| 6 | ↓ 1 | claude-opus-4-6 | Anthropic | 1548 ±5 | 23351 | $5 / $25 | 1M |
| 7 | ↓ 1 | claude-opus-4-7 | Anthropic | 1546 ±6 | 18816 | $5 / $25 | 1M |
| 8 | - | gpt-6.1-sol-max | OpenAI | 1545 ±20 | 911 | $2 / $10 | 1.05M |
| 9 | ↓ 2 | gpt-6-astra-max | OpenAI | 1542 ±12 | 2424 | N/A | N/A |
| 10 | ↓ 1 | kimi-k3-max | Moonshot | 1540 ±7 | 7775 | N/A | N/A |
| — Below are domestic models outside the Top 10 — | |||||||
| 11 | ↓ 1 | mimo-v2.6-pro | Xiaomi | 1539 ±18 | 1107 | $0.44 / $0.87 | 1.05M |
| 26 | - | deepseek-v4.1-flash-max | DeepSeek | 1528 ±11 | 3026 | $0.3 / $1.2 | 1M |
| 27 | ↑ 2 | glm-5.3-flash | Z.ai | 1525 ±8 | 7267 | $0.06 / $0.2 | 1.05M |
| 28 | ↓ 1 | qwen3.7-max-preview | Alibaba | 1524 ±18 | 1165 | $1.48 / $4.43 | 1M |
| 29 | ↓ 1 | qwen3.8-max | Alibaba | 1524 ±8 | 7077 | $2 / $6 | 1M |
Frontend Development Leaderboard
The top of the frontend development leaderboard remains stable, with claude-opus-5.5-max continuing to rank first at an ELO of 1813, holding a clear lead. Among domestic models, Xiaomi's mimo-v2.6-flash newly entered the leaderboard this week at rank 16 with an ELO of 1637; mimo-v2.6-pro climbed 5 places to 19th with an ELO of 1629, a remarkable performance. Alibaba's qwen3.8-max-0902 rose 1 place to 10th, and qwen3.8-max sits at 11th, with both models entering the top 15, showing that domestic frontend development capability continues to improve.
| Rank | vs. last week | Model | Vendor | Score (±CI) | Votes | Price ($/M) Input / Output | Context |
|---|---|---|---|---|---|---|---|
| 1 | - | claude-opus-5.5-max | Anthropic | 1813 ±15 | 2294 | $4 / $20 | 1M |
| 2 | - | gpt-6-astra-max | OpenAI | 1786 ±11 | 5189 | $10 / $50 | 1.05M |
| 3 | - | claude-sonnet-5.5-xhigh | Anthropic | 1774 ±15 | 2323 | $2 / $10 | 1M |
| 4 | - | gpt-6.1-sol-max | OpenAI | 1755 ±15 | 2102 | $2 / $10 | 1.05M |
| 5 | - | claude-fable-5.1-max | Anthropic | 1744 ±9 | 6663 | $10 / $50 | 1M |
| 6 | - | claude-sonnet-5.5-high | Anthropic | 1715 ±11 | 3984 | $2 / $10 | 1M |
| 7 | - | claude-opus-5-max | Anthropic | 1691 ±6 | 17320 | $5 / $25 | 1M |
| 8 | - | gpt-6-sol-max | OpenAI | 1688 ±12 | 3407 | $2 / $10 | 1.05M |
| 9 | - | gemini-4-argon-high | 1678 ±15 | 1771 | $2 / $10 | 1M | |
| 10 | ↑ 1 | qwen3.8-max-0902 | Alibaba | 1674 ±8 | 11182 | $2 / $6 | N/A |
| — Below are domestic models outside the Top 10 — | |||||||
| 11 | ↓ 1 | qwen3.8-max | Alibaba | 1672 ±12 | 3446 | $2 / $6 | 1M |
| 14 | ↓ 1 | kimi-k3-max | Moonshot | 1654 ±6 | 17412 | $3 / $15 | 1.05M |
| 16 | Newly listed | mimo-v2.6-flash | Xiaomi | 1637 ±10 | 4058 | $0.14 / $0.28 | 1.05M |
| 17 | ↓ 1 | qwen3.8-flash-next | Alibaba | 1633 ±8 | 6932 | $0.16 / $0.47 | 1M |
| 18 | ↓ 1 | hy4-preview | Tencent | 1632 ±10 | 4792 | $0.83 / $2.5 | 1.05M |
| 22 | ↓ 2 | glm-5.3-max | Z.ai | 1622 ±8 | 8382 | $1.4 / $4.4 | 1.05M |
| 23 | ↓ 2 | deepseek-v4.1-flash-max | DeepSeek | 1619 ±11 | 3542 | $0.3 / $1.2 | 1M |
| 26 | ↓ 1 | glm-5.3-flash | Z.ai | 1609 ±7 | 13697 | $0.06 / $0.2 | 1.05M |
| 27 | ↓ 1 | glm-5.2-max | Z.ai | 1603 ±6 | 15351 | $1.4 / $4.4 | 1M |
| 28 | - | qwen3.8-27b | Alibaba | 1593 ±7 | 13777 | $0.4 / $3 | N/A |
Agent Leaderboard
The agent leaderboard sees a major change this edition: Anthropic Claude Opus 5.5 (Max) debuts and immediately takes the top spot with a net improvement of 15.53%, a task confirmation success rate of 17.29%, and a tool hallucination rate of just 0.37%, with all signal metrics at the leading level. Claude Opus 5.5 (High) holds second place with a net improvement of 14.33%, while GPT 6 Astra (Max) rises 1 place to third. Gemini 4 Argon (High) climbs 2 places to eighth, and its tool hallucination rate of just 0.18% is the lowest among current top models.
Domestic models overall sit in the second tier of the agent leaderboard; specific performances:
DeepSeek Deepseek V4.1 Flash (Max) ranks 17th, with a net improvement of 3.80%, a task confirmation success rate of 7.44%, unchanged in rank;
Moonshot AI Kimi K3 (Max) ranks 18th, with a net improvement of 3.78%, a task confirmation success rate of 6.46%, down 4 places from last week;
Tencent Hy4 preview ranks 19th, with a net improvement of 3.61%, a task confirmation success rate of 5.96%, unchanged in rank;
Xiaomi MiMo V2.6 Pro ranks 20th, with a net improvement of 3.52%, a task confirmation success rate of 8.03%, up 1 place from last week;
Alibaba Qwen3.8 Max ranks 24th, with a net improvement of 2.35%, a task confirmation success rate of 5.09%, down 1 place from last week.
| Rank | vs. last week | Model | Vendor | Net Improvement (±CI) | Task Confirmation Success Rate | Positive / Negative Review Ratio | Controllability |
|---|---|---|---|---|---|---|---|
| 1 | New Entry | Claude Opus 5.5 (Max) | Anthropic | 15.53% ±4.63% | 17.29% ±6.02% | 36.78% ±20.55% | 8.2% ±4.06% |
| 2 | - | Claude Opus 5.5 (High) | Anthropic | 14.33% ±2.18% | 13.93% ±3.84% | 33.69% ±9.12% | 9.98% ±2.12% |
| 3 | ↑ 1 | GPT 6 Astra (Max) | OpenAI | 13.09% ±2.37% | 12.13% ±3.88% | 38.54% ±9.96% | 8.85% ±2.5% |
| 4 | ↓ 3 | Claude Fable 5.1 (Max) | Anthropic | 12.66% ±2.05% | 16.19% ±2.96% | 26.22% ±7.98% | 8.84% ±3.12% |
| 5 | ↓ 2 | Claude Sonnet 5.5 (Max) | Anthropic | 11.95% ±3.31% | 15.83% ±5.06% | 23.37% ±14.09% | 5.1% ±3.33% |
| 6 | ↓ 1 | GPT 6.1 Sol (Max) | OpenAI | 11.72% ±3.05% | 14.74% ±4.18% | 30.32% ±12.99% | 8.52% ±3.07% |
| 7 | ↓ 1 | GPT 6 Sol (Max) | OpenAI | 9.92% ±2.38% | 6.97% ±4.34% | 23.55% ±9.64% | 11.92% ±2.02% |
| 8 | ↑ 2 | Gemini 4 Argon (High) | 9.26% ±2.36% | 15.64% ±3.26% | 40.6% ±9.56% | 11.64% ±2.1% | |
| 9 | ↓ 1 | Claude Fable 5 (High) | Anthropic | 8.94% ±1.27% | 5.79% ±2.6% | 19.34% ±4.8% | 9.49% ±1.39% |
| 10 | ↓ 1 | Claude Opus 5 (Max) | Anthropic | 8.12% ±1.47% | 10.42% ±3.1% | 11.86% ±5.46% | 4.22% ±2.01% |
| — The following are domestic (Chinese) models outside the Top 10 — | |||||||
| 17 | - | Deepseek V4.1 Flash (Max) | DeepSeek | 3.8% ±0.39% | 7.44% ±0.86% | 2.44% ±1.45% | 0.87% ±0.54% |
| 18 | ↓ 4 | Kimi K3 (Max) | Moonshot | 3.78% ±0.59% | 6.46% ±1.32% | 6.44% ±2.05% | 1.14% ±0.86% |
| 19 | - | Hy4 preview | Tencent | 3.61% ±0.7% | 5.96% ±1.48% | 5.18% ±2.59% | 0.32% ±0.92% |
| 20 | ↑ 1 | MiMo V2.6 Pro | Xiaomi | 3.52% ±1.92% | 8.03% ±3.88% | 8.16% ±7.31% | 0.56% ±2.45% |
| 21 | ↓ 1 | GLM 5.2 (Max) | Z.ai | 3.19% ±0.73% | 4.11% ±1.58% | 8.94% ±2.61% | 1.72% ±0.96% |
| 24 | ↓ 1 | Qwen3.8 Max | Alibaba | 2.35% ±0.59% | 5.09% ±1.32% | 2.95% ±2.13% | 1.81% ±0.82% |
| 25 | ↓ 1 | GLM 5.3 (Max) | Z.ai | 2.15% ±0.72% | 4.78% ±1.62% | 5.12% ±2.48% | 0.44% ±1.08% |
| 27 | ↑ 1 | DeepSeek V4 Pro (High) (0813) | DeepSeek | 1.34% ±1.66% | 2% ±4.01% | 1.86% ±5.99% | 2.12% ±2.2% |
AI Image Generation Leaderboard
The top of the AI image generation leaderboard is dominated by OpenAI models, with gpt-image-2.5-sunburst continuing to hold first place firmly at an ELO of 1425, leading the second place by more than 20 points. Google's gemini-nano-banana-2.1 newly entered the leaderboard this week, debuting at sixth place with an ELO of 1328, a decent performance. Among domestic models, ByteDance's seedream-5.0-pro ranks 12th and Alibaba's qwen-image-3.0-pro ranks 13th, both in the upper-middle tier, with little change in rankings.
| Rank | vs. last week | Model | Provider | Score (±CI) | Votes | Price ($/M) Input / Output | Context |
|---|---|---|---|---|---|---|---|
| 1 | - | gpt-image-2.5-sunburst | OpenAI | 1425 ±7 | 17752 | N/A | N/A |
| 2 | - | gpt-image-2.5-flare | OpenAI | 1398 ±7 | 16889 | N/A | N/A |
| 3 | - | gpt-image-2 (medium) | OpenAI | 1383 ±4 | 93334 | $8 / $30 | N/A |
| 4 | ↑ 1 | grok-imagine-image-2.0 (canvas) | SpaceXAI | 1334 ±12 | 3113 | N/A | N/A |
| 5 | ↓ 1 | mai-image-2.6 | Microsoft AI | 1332 ±5 | 25091 | N/A | N/A |
| 6 | New entry | gemini-nano-banana-2.1 | 1328 ±9 | 5310 | $1.5 / $7.5 | 65.5K | |
| 7 | ↓ 1 | reve-2.1 | Reve | 1302 ±8 | 8249 | N/A | N/A |
| 8 | ↓ 1 | grok-imagine-image-2.0 (20260801) | SpaceXAI | 1294 ±6 | 14483 | N/A | N/A |
| 9 | ↓ 1 | muse-image | Meta | 1274 ±5 | 42262 | N/A | N/A |
| 10 | ↓ 1 | reve-2.0 | Reve | 1269 ±6 | 15773 | N/A | N/A |
| — Below are the domestic Chinese models outside the Top 10 — | |||||||
| 12 | ↓ 1 | seedream-5.0-pro | Bytedance | 1255 ±4 | 94181 | N/A | N/A |
| 13 | ↓ 1 | qwen-image-3.0-pro | Alibaba | 1255 ±6 | 11440 | N/A | N/A |
| 19 | ↓ 1 | qwen-image-2.1 | Alibaba | 1223 ±7 | 11441 | N/A | N/A |
| 21 | ↓ 1 | qwen-image-2.0-pro-2026-06-22 | Alibaba | 1192 ±6 | 13096 | N/A | N/A |
AI Video Leaderboard
The top of the AI video leaderboard remains dominated by Google models, with gemini-omni-1.1-flash and gemini-omni-flash taking the top two spots respectively, both with ELO scores exceeding 1510. Domestic (Chinese) models performed notably well, with 6 models entering the top 20: ByteDance's dreamina-seedance-2.0-720p ranked fifth, Alibaba's wan3.0 ranked sixth, dreamina-seedance-2.5-720p ranked seventh, minimax-h3 ranked eighth, and Alibaba's happyhorse-1.0 ranked tenth, all holding stable positions, reflecting the overall competitiveness of Chinese AI video generation models.
| Rank | vs last week | Model | Vendor | Score (±CI) | Votes | Price ($/M) Input / Output | Context |
|---|---|---|---|---|---|---|---|
| 1 | - | gemini-omni-1.1-flash | 1516 ±15 | 1784 | $1.5 / $9 | N/A | |
| 2 | - | gemini-omni-flash | 1513 ±9 | 26576 | N/A | N/A | |
| 3 | - | flux-3-video | Black Forest Labs | 1493 ±17 | 1302 | N/A | N/A |
| 4 | - | grok-imagine-video-1.5-agent | SpaceXAI | 1492 ±18 | 1218 | N/A | N/A |
| 5 | - | dreamina-seedance-2.0-720p | Bytedance | 1479 ±8 | 56318 | N/A | N/A |
| 6 | - | wan3.0 | Alibaba | 1476 ±13 | 2598 | N/A | N/A |
| 7 | - | dreamina-seedance-2.5-720p | Bytedance | 1474 ±9 | 8001 | N/A | N/A |
| 8 | - | minimax-h3 | MiniMax | 1460 ±9 | 11417 | N/A | N/A |
| 9 | - | muse-video | Meta | 1456 ±15 | 2188 | N/A | N/A |
| 10 | - | happyhorse-1.0 | Alibaba-ATH | 1427 ±13 | 22264 | N/A | N/A |
| — Below are the Chinese models outside the Top 10 — | |||||||
| 20 | - | wan2.7-t2v | Alibaba | 1337 ±7 | 27626 | N/A | N/A |
| 21 | - | wan2.6-t2v | Alibaba | 1327 ±7 | 54050 | N/A | N/A |
| 22 | - | seedance-v1.5-pro | Bytedance | 1256 ±7 | 76171 | N/A | N/A |
| 25 | - | wan2.5-t2v-preview | Alibaba | 1245 ±9 | 36378 | N/A | N/A |
Looking at this week's Arena leaderboard as a whole, Anthropic's new models performed strongly: Claude Opus 5.5 Max made its debut at the top of the agentic leaderboard, and Claude models also held top positions across multiple sub-leaderboards, further tilting the competitive landscape toward Anthropic. Domestic (Chinese) models remained stable overall, with a clear multi-tier distribution: kimi-k3-max, qwen3.8-max and others continued to hold steady in the upper-middle ranks, while the new model mimo-v2.6-flash also successfully entered the leaderboard, reflecting the steady iteration pace of domestic models. More domestic large model version updates are expected next week, so ranking changes are worth continued attention.