# Best LLMs for Writing in 2026

> AI models ranked for writing quality — instruction following, tone control, and human preference scores.

There is no single "writing benchmark," so BenchLM ranks writing capability by the instruction-following category — the best available proxy for whether a model matches your tone, structure, and length constraints — read alongside human-preference ratings from an independent index. Claude models held the top of that index in the September 2026 snapshot; the instruction-following table shows who follows a brief most reliably.

Canonical page: https://benchlm.ai/best/writing

Last updated: October 6, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score |
|------|-------|---------|------|---------|-------|
| 1 | [MAI-Thinking-1](/models/mai-thinking-1) | Microsoft | Proprietary | 256K | 94.7 |
| 2 | [Grok 4.3](/models/grok-4-3) | xAI | Proprietary | 1M | 93.2 |
| 3 | [GPT-5.2-Codex](/models/gpt-5-2-codex) | OpenAI | Proprietary | 400K | 92.4 |
| 4 | [GLM-5.1](/models/glm-5-1) | Z.AI | Open Weight | 203K | 92.4 |
| 5 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 92.4 |
| 6 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | Xiaomi | Proprietary | 1M | 92.4 |
| 7 | [GPT-5.5](/models/gpt-5-5) | OpenAI | Proprietary | 1M | 91.9 |
| 8 | [Muse Spark](/models/muse-spark) | Meta | Proprietary | 262K | 91.9 |
| 9 | [GPT-5.4 nano](/models/gpt-5-4-nano) | OpenAI | Proprietary | 400K | 91.9 |
| 10 | [MiniMax M2.7](/models/minimax-m2-7) | MiniMax | Open Weight | 200K | 91.6 |
| 11 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 91.6 |
| 12 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 91.5 |
| 13 | [Gemma 4 31B](/models/gemma-4-31b) | Google | Open Weight | 256K | 91.5 |
| 14 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | OpenAI | Proprietary | 400K | 91.2 |
| 15 | [GPT-5.2](/models/gpt-5-2) | OpenAI | Proprietary | 400K | 91.2 |
| 16 | [Qwen3.8 Max](/models/qwen3-8-max) | Alibaba | Open Weight | 1M | 89.8 |
| 17 | [Qwen3.7 Max](/models/qwen3-7-max) | Alibaba | Proprietary | 1M | 89.2 |
| 18 | [Qwen3.7 Plus](/models/qwen3-7-plus) | Alibaba | Proprietary | 1M | 89.2 |
| 19 | [GPT-5.4](/models/gpt-5-4) | OpenAI | Proprietary | 1.05M | 89.2 |
| 20 | [Command A+](/models/command-a-plus) | Cohere | Open Weight | 128K | 89.2 |
| 21 | [Gemma 4 12B](/models/gemma-4-12b) | Google | Open Weight | 256K | 88.7 |
| 22 | [GLM-5.2](/models/glm-5-2) | Z.AI | Open Weight | 1M | 88.5 |
| 23 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | Open Weight | 1M | 88.5 |
| 24 | [GPT-5.4 mini](/models/gpt-5-4-mini) | OpenAI | Proprietary | 400K | 88.5 |
| 25 | [GLM-5-Turbo](/models/glm-5-turbo) | Z.AI | Proprietary | 200K | 88.3 |
| 26 | [GPT-5.1](/models/gpt-5-1) | OpenAI | Proprietary | 200K | 87.9 |
| 27 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | Proprietary | 1.05M | 87.7 |
| 28 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | NVIDIA | Open Weight | 1M | 87.4 |
| 29 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 87.4 |
| 30 | [Gemma 4 26B A4B](/models/gemma-4-26b-a4b) | Google | Open Weight | 256K | 87.3 |
| 31 | [GLM-5](/models/glm-5) | Z.AI | Open Weight | 200K | 87.2 |
| 32 | [Qwen3.8-Omni-Flash](/models/qwen3-8-omni-flash) | Alibaba | Proprietary | 1M | 86.9 |
| 33 | [Qwen3.8-Flash-Next](/models/qwen3-8-flash-next) | Alibaba | Open Weight | 262K | 86.5 |
| 34 | [o3](/models/o3) | OpenAI | Proprietary | 200K | 86 |
| 35 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | Proprietary | 1.05M | 85.7 |
| 36 | [Solar Pro 3](/models/solar-pro-3) | Upstage | Proprietary | 128K | 85.7 |
| 37 | [GPT-5 (medium)](/models/gpt-5-medium) | OpenAI | Proprietary | 128K | 84.9 |
| 38 | [Gemini 3 Pro](/models/gemini-3-pro) | Google | Proprietary | 2M | 84.7 |
| 39 | [dots3-note Preview](/models/dots3-note-preview) | Dots Studio | Open Weight | 512K | 84.5 |
| 40 | [o1](/models/o1) | OpenAI | Proprietary | 200K | 84.5 |
| 41 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 84.4 |
| 42 | [GPT-5.1-Codex](/models/gpt-5-1-codex) | OpenAI | Proprietary | 400K | 84.2 |
| 43 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | Proprietary | 1M | 84.1 |
| 44 | [Inkling](/models/inkling) | Thinking Machines Lab | Open Weight | 1M | 83.2 |
| 45 | [GPT-OSS 120B](/models/gpt-oss-120b) | OpenAI | Open Weight | 128K | 82.8 |
| 46 | [MiMo-V2-Pro](/models/mimo-v2-pro) | Xiaomi | Proprietary | 1M | 82.6 |
| 47 | [Mistral Medium 3.5 128B](/models/mistral-medium-3-5-128b) | Mistral | Open Weight | 256K | 82.6 |
| 48 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | Proprietary | 1M | 82.5 |
| 49 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | Open Weight | 262K | 82.5 |
| 50 | [Granite 4.2 8B](/models/granite-4-2-8b) | IBM | Open Weight | 128K | 82.1 |
| 51 | [GLM-4.7](/models/glm-4-7) | Z.AI | Open Weight | 200K | 81.4 |
| 52 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | Open Weight | 262K | 81 |
| 53 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | Open Weight | 256K | 80.6 |
| 54 | [GPT-OSS 20B](/models/gpt-oss-20b) | OpenAI | Open Weight | 128K | 77.8 |
| 55 | [Muse Glimmer 30B](/models/muse-glimmer-30b) | Meta | Open Weight | 131K | 77 |
| 56 | [Mercury 2.5](/models/mercury-2-5) | Inception | Proprietary | 260K | 77 |
| 57 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | Open Weight | 262K | 76.8 |
| 58 | [Claude Fable 5](/models/claude-fable) | Anthropic | Proprietary | 1M+ | 75.7 |
| 59 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | Moonshot AI | Open Weight | 256K | 75.1 |
| 60 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | Proprietary | 1M | 74 |
| 61 | [GLM-5V-Turbo](/models/glm-5v-turbo) | Z.AI | Proprietary | 200K | 72.5 |
| 62 | [Nemotron 3 Nano Omni 30B A3B](/models/nemotron-3-nano-omni-30b-a3b) | NVIDIA | Open Weight | 256K | 72.4 |
| 63 | [Ling 3.0 Flash](/models/ling-3-0-flash) | InclusionAI | Open Weight | 262K | 71.4 |
| 64 | [Ternary Bonsai 2 27B](/models/ternary-bonsai-2-27b) | Prism ML | Open Weight | 262K | 70.3 |
| 65 | [Ling 3.0 Flash FP8](/models/ling-3-0-flash-fp8) | InclusionAI | Open Weight | 262K | 69 |
| 66 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | Anthropic | Proprietary | 200K | 68.5 |
| 67 | [Nemotron 3.5 Lightning 30B A3B NVFP4](/models/nemotron-3-5-lightning-30b-a3b-nvfp4) | NVIDIA | Open Weight | 1M | 67.8 |
| 68 | [Trinity-Large-Thinking](/models/trinity-large-thinking) | Arcee AI | Open Weight | 512K | 66.3 |
| 69 | [Claude 4.1 Opus Thinking](/models/claude-4-1-opus-thinking) | Anthropic | Proprietary | 200K | 65.1 |
| 70 | [Gemini 3 Flash](/models/gemini-3-flash) | Google | Proprietary | 1M | 64.7 |
| 71 | [Grok 4](/models/grok-4) | xAI | Proprietary | 128K | 62.9 |
| 72 | [MiMo-V2-Omni](/models/mimo-v2-omni) | Xiaomi | Proprietary | 262K | 62.6 |
| 73 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | xAI | Proprietary | 2M | 61.6 |
| 74 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | xAI | Proprietary | 2M | 58.7 |
| 75 | [DeepSeek V3.2](/models/deepseek-v3-2) | DeepSeek | Open Weight | 128K | 56.7 |
| 76 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | Google | Proprietary | 1M | 56.4 |
| 77 | [Mistral Small 4](/models/mistral-small-4) | Mistral | Open Weight | 256K | 55.7 |
| 78 | [Claude 4 Sonnet](/models/claude-4-sonnet) | Anthropic | Proprietary | 200K | 52 |
| 79 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | Proprietary | 1M | 51 |
| 80 | [Gemma 4 E4B](/models/gemma-4-e4b) | Google | Open Weight | 128K | 50.5 |
| 81 | [Qwen3 Max](/models/qwen3-max) | Alibaba | Proprietary | 1M | 50.3 |
| 82 | [GPT-4.1](/models/gpt-4-1) | OpenAI | Proprietary | 1M | 48.9 |
| 83 | [Llama 4 Maverick](/models/llama-4-maverick) | Meta | Open Weight | 1M | 48.9 |
| 84 | [DeepSeek V3.1 (Reasoning)](/models/deepseek-v3-1-reasoning) | DeepSeek | Open Weight | 128K | 47 |
| 85 | [Kimi K2](/models/kimi-k2) | Moonshot AI | Proprietary | 128K | 47 |
| 86 | [Grok Code Fast 1](/models/grok-code-fast-1) | xAI | Proprietary | 256K | 46.8 |
| 87 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | Proprietary | 200K | 46.6 |
| 88 | [DeepSeek V3 0324](/models/deepseek-v3-0324) | DeepSeek | Open Weight | 128K | 46.3 |
| 89 | [Hy3 Preview](/models/hy3-preview) | Tencent | Open Weight | 256K | 46.1 |
| 90 | [MiMo-V2-Flash](/models/mimo-v2-flash) | Xiaomi | Open Weight | 256K | 44.9 |
| 91 | [DeepSeek-R1](/models/deepseek-r1) | DeepSeek | Open Weight | 128K | 44.5 |
| 92 | [Llama 4 Scout](/models/llama-4-scout) | Meta | Open Weight | 10M | 44.3 |
| 93 | [Mistral Medium 3](/models/mistral-medium-3) | Mistral | Proprietary | 128K | 44.1 |
| 94 | [Gemini 2.5 Flash](/models/gemini-2-5-flash) | Google | Proprietary | 1M | 43.7 |
| 95 | [GPT-4.1 mini](/models/gpt-4-1-mini) | OpenAI | Proprietary | 1M | 42.8 |
| 96 | [Nova Pro](/models/nova-pro) | Amazon | Proprietary | 128K | 42.5 |
| 97 | [Gemma 4 E2B](/models/gemma-4-e2b) | Google | Open Weight | 128K | 42.4 |
| 98 | [DeepSeek V3.1](/models/deepseek-v3-1) | DeepSeek | Open Weight | 128K | 42.1 |
| 99 | [GLM-4.5-Air](/models/glm-4-5-air) | Z.AI | Proprietary | 128K | 41.9 |
| 100 | [GLM-4.6](/models/glm-4-6) | Z.AI | Open Weight | 200K | 40.7 |
| 101 | [Grok 4.1 Fast](/models/grok-4-1-fast) | xAI | Proprietary | 1M | 40.4 |
| 102 | [Mistral Large 3](/models/mistral-large-3) | Mistral | Proprietary | 128K | 40 |
| 103 | [Claude 3 Haiku](/models/claude-3-haiku) | Anthropic | Proprietary | 200K | 39.9 |
| 104 | [DeepSeek V3](/models/deepseek-v3) | DeepSeek | Open Weight | 128K | 38.2 |
| 105 | [Sarvam 105B](/models/sarvam-105b) | Sarvam | Open Weight | 128K | 37.7 |
| 106 | [GPT-4o](/models/gpt-4o) | OpenAI | Proprietary | 128K | 37.6 |
| 107 | [Ling 2.6 Flash](/models/ling-2-6-flash) | InclusionAI | Open Weight | 262K | 37.5 |
| 108 | [LFM2.5-2.6B](/models/lfm2-5-2-6b) | LiquidAI | Open Weight | 128K | 37.4 |
| 109 | [LFM2.5-8B-A1B](/models/lfm2-5-8b-a1b) | LiquidAI | Open Weight | 128K | 35.4 |
| 110 | [GPT-4.1 nano](/models/gpt-4-1-nano) | OpenAI | Proprietary | 1M | 34.6 |
| 111 | [Gemma 3 27B](/models/gemma-3-27b) | Google | Open Weight | 32K | 34.3 |
| 112 | [Mistral Large 2](/models/mistral-large-2) | Mistral | Proprietary | 128K | 33.5 |
| 113 | [Qwen3-Omni-30B-A3B-Instruct](/models/qwen3-omni-30b-a3b-instruct) | Alibaba | Open Weight | N/A | 33.5 |
| 114 | [GPT-4o mini](/models/gpt-4o-mini) | OpenAI | Proprietary | 128K | 33.2 |
| 115 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | Proprietary | 200K | 30.6 |
| 116 | [DeepSeek R1 Distill Qwen 32B](/models/deepseek-r1-distill-qwen-32b) | DeepSeek | Open Weight | 128K | 27.4 |
| 117 | [Granite-4.0-H-1B](/models/granite-4-0-h-1b) | IBM | Open Weight | 128K | 27.4 |
| 118 | [Sarvam 30B](/models/sarvam-30b) | Sarvam | Open Weight | 64K | 27.4 |
| 119 | [Granite-4.0-350M](/models/granite-4-0-350m) | IBM | Open Weight | 32K | 27.4 |
| 120 | [Granite-4.0-H-350M](/models/granite-4-0-h-350m) | IBM | Open Weight | 32K | 27.4 |
| 121 | [Phi-4](/models/phi-4) | Microsoft | Open Weight | 16K | 27.4 |
| 122 | [ZAYA1-8B](/models/zaya1-8b) | Zyphra | Open Weight | 131K | 22.8 |
| 123 | [MiniCPM5-1B](/models/minicpm5-1b) | OpenBMB | Open Weight | 131K | 9.7 |
| 124 | [LFM2.5-230M](/models/lfm2-5-230m) | LiquidAI | Open Weight | 32K | 9.7 |

## Key Takeaways

- Top model: [MAI-Thinking-1](/models/mai-thinking-1) with a score of 94.7
- Best open-weight option: [GLM-5.1](/models/glm-5-1) at #4
- Models included: 124

## Compare the Leaders

- [MAI-Thinking-1 vs Grok 4.3](/compare/grok-4-3-vs-mai-thinking-1)