# Best AI Models for Web Research in 2026

> AI models ranked on sourced browsing and research benchmarks including BrowseComp, WebArena, and WebVoyager.

This reporting page isolates the web research slice of agentic performance. It prioritizes sourced benchmarks for browsing, evidence gathering, and multi-step web task completion rather than generic overall agent scores.

Canonical page: https://benchlm.ai/best/web-research

Last updated: October 5, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score |
|------|-------|---------|------|---------|-------|
| 1 | [Atria Dawn Preview](/models/atria-dawn-preview) | Shanghai Artificial Intelligence Laboratory | Open Weight | 256K | 92.5 |
| 2 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | Proprietary | 1.05M | 92.2 |
| 3 | [GPT-6 Astra](/models/gpt-6-astra) | OpenAI | Proprietary | 1.05M | 91.5 |
| 4 | [Kimi K3](/models/kimi-k3) | Moonshot AI | Pending | 1.05M | 91.2 |
| 5 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | Proprietary | null | 90.8 |
| 6 | [GPT-5.5 Pro](/models/gpt-5-5-pro) | OpenAI | Proprietary | 1M | 90.1 |
| 7 | [Fara1.5-27B](/models/fara-1-5-27b) | Microsoft | Open Weight | 262K | 89.3 |
| 8 | [GPT-5.4 Pro](/models/gpt-5-4-pro) | OpenAI | Proprietary | 1.05M | 89.3 |
| 9 | [Step 5 Preview](/models/step-5-preview) | StepFun | Pending | 1M | 88.7 |
| 10 | [Claude Mythos 5](/models/claude-mythos-5) | Anthropic | Proprietary | 1M+ | 88 |
| 11 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | Proprietary | 1.05M | 87.5 |
| 12 | [Ornith-1.5-397B](/models/ornith-1-5-397b) | Ornith AI | Open Weight | 262K | 86.6 |
| 13 | [Claude Sonnet 5](/models/claude-sonnet-5) | Anthropic | Proprietary | 1M | 84.7 |
| 14 | [GPT-5.5](/models/gpt-5-5) | OpenAI | Proprietary | 1M | 84.4 |
| 15 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | Proprietary | 1M | 84.3 |
| 16 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | Proprietary | 1M | 83.7 |
| 17 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 83.5 |
| 18 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | DeepSeek | Open Weight | 1M | 83.4 |
| 19 | [dots3-note Preview](/models/dots3-note-preview) | Dots Studio | Open Weight | 512K | 83.3 |
| 20 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | Proprietary | 1.05M | 83.3 |
| 21 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | Open Weight | 256K | 83.2 |
| 22 | [GPT-5.4](/models/gpt-5-4) | OpenAI | Proprietary | 1.05M | 82.7 |
| 23 | [Fara1.5-4B](/models/fara-1-5-4b) | Microsoft | Open Weight | 262K | 80.8 |
| 24 | [Claude Opus 4.7 (Adaptive)](/models/claude-opus-4-7-adaptive) | Anthropic | Proprietary | 1M | 79.3 |
| 25 | [Beam](/models/reflection-beam) | Reflection AI | Pending | null | 77.4 |
| 26 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | Open Weight | 1M | 77.4 |
| 27 | [Inkling](/models/inkling) | Thinking Machines Lab | Open Weight | 1M | 77.1 |
| 28 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | Open Weight | 256K | 75.8 |
| 29 | [Agents-A1](/models/agents-a1) | InternScience | Open Weight | 262K | 75.5 |
| 30 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | DeepSeek | Open Weight | 1M | 73.2 |
| 31 | [Ling 3.0 Flash](/models/ling-3-0-flash) | InclusionAI | Open Weight | 262K | 72.2 |
| 32 | [GLM-5.1](/models/glm-5-1) | Z.AI | Open Weight | 203K | 68 |
| 33 | [Ornith-1.5-35B-A3B](/models/ornith-1-5-35b-a3b) | Ornith AI | Open Weight | 262K | 67.6 |
| 34 | [Agents-A1-4B](/models/agents-a1-4b) | InternScience | Open Weight | 262K | 66.8 |
| 35 | [GPT-5.2](/models/gpt-5-2) | OpenAI | Proprietary | 400K | 65.8 |
| 36 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 63.8 |
| 37 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 62 |
| 38 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 61 |
| 39 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 61 |
| 40 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 60.6 |
| 41 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | Moonshot AI | Proprietary | 128K | 60.6 |
| 42 | [Ornith-1.5-9B](/models/ornith-1-5-9b) | Ornith AI | Open Weight | 262K | 56.4 |
| 43 | [GLM-4.7](/models/glm-4-7) | Z.AI | Open Weight | 200K | 52 |
| 44 | [Solar Pro 4](/models/solar-pro-4) | Upstage | Proprietary | 512K | 49.2 |
| 45 | [LongCat-Flash-Lite-Sparse](/models/longcat-flash-lite-sparse) | Meituan | Open Weight | 1M | 48.6 |
| 46 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | NVIDIA | Open Weight | 1M | 44.4 |
| 47 | [Nemotron 3.5 Lightning 30B A3B NVFP4](/models/nemotron-3-5-lightning-30b-a3b-nvfp4) | NVIDIA | Open Weight | 1M | 36.8 |

## Key Takeaways

- Top model: [Atria Dawn Preview](/models/atria-dawn-preview) with a score of 92.5
- Best open-weight option: [Atria Dawn Preview](/models/atria-dawn-preview) at #1
- Models included: 47

## Compare the Leaders

- [Atria Dawn Preview vs GPT-5.6 Sol](/compare/atria-dawn-preview-vs-gpt-5-6-sol)