Best LLMs for Writing (2026)

Last refreshed 2026-10-17. Next refresh: weekly.

The best LLMs for writing in 2026, ranked by human preference. Covers long-form essays, creative prose, and marketing copy — with pricing.

Verdict

Use Claude Opus 4.7 for long-form writing today.

Claude Opus 4.6 is the runner-up: 1503 vs 1501 on Arena.

Researched 1d agoWhy this pickMethodology

How we rank

Writing picks are for essays, drafts, and long-form prose. The target rubric is EQ-Bench Creative Writing v3 when coverage is available; the live fallback is Chatbot Arena, then MMLU.

  1. Eligibility — General chat models excluding pure code/embedding SKUs.
  2. Target ranking — EQ-Bench Creative Writing v3 becomes the primary signal once it exists in seed data for enough current models.
  3. Current live ranking — Until that coverage lands, the page keeps the current Chatbot Arena preference score fallback, then MMLU, then newer release.
  4. Variant collapse — We keep one row per model family (`familySlug` + parameter tier). When headline scores tie within ±0.5 pt (±10 Elo on Chatbot Arena), we pick the canonical SKU by lowest tracked input price, then GA over preview or limited access, then newest `release`. A folded sibling within the benchmark noise band can show a "Tied within margin" chip on that score cell.
  5. Why this ranking? — Arena is an imperfect but useful human-preference proxy for prose quality; MMLU is only a general capability floor, not a style or brand-voice metric.
  6. Marketing boundary — Use `/best/marketing` for ad copy, email, social posts, localization, and brand-voice content; this page stays focused on general writing.
Chatbot ArenaMMLU
#ModelInput $/1MOutput $/1M
1Claude Opus 4.7
ReasoningVisionTools

Arena: 1503

$5.00$25.00
2Claude Opus 4.6
ReasoningVisionTools

Arena: 1501

$5.00$25.00
3Gemini 3.1 Pro Preview
PreviewVisionTools

Arena: 1493

$2.00$12.00
4Muse Spark
ReasoningVisionTools

Arena: 1491

——
5GPT-5.5
ReasoningVisionTools

Arena: 1488

$5.00$30.00
6GPT-5.4
ReasoningVisionTools

Arena: 1479

$2.50$15.00
7ERNIE 5.1
Tools

Arena: 1476

$0.59$2.65
8Qwen3.7-Max
ReasoningTools

Arena: 1475

$1.25$3.75
9GLM-5.1
ReasoningTools

Arena: 1475

$1.38$4.40
10Gemini 3 Flash
PreviewVisionTools

Arena: 1467

$0.50$3.00
11Claude Opus 4.5
ReasoningVisionTools

Arena: 1466

$5.00$25.00
12Grok 4.1
ReasoningVisionTools

Arena: 1464

——
13Claude Sonnet 4.6
ReasoningVisionTools

Arena: 1459

$3.00$15.00
14DeepSeek V4 Pro
ReasoningTools

Arena: 1456

$0.43$0.87
15Gemini 3.1 Flash-Lite
VisionTools

Arena: 1432

$0.25$1.50
16o3
ReasoningVisionTools

Arena: 1412

$2.00$8.00
17Gemini 2.5 Pro
ReasoningVisionTools

Arena: 1398

$1.25$10.00
18DeepSeek R1
Reasoning

Arena: 1372

$0.10$0.30
19Llama 4 Maverick 17B Instruct FP8
Vision

Arena: 1365

$0.15$0.60
20Gemini 2.5 Flash
VisionTools

Arena: 1320

$0.30$2.50

Honorable mentions

Next seats in this ranking. Lines below are from each model's stored description in LLMReference seed data—spot-check the model page before relying on a capability claim.
  • #4ERNIE 5.1

    ERNIE 5.1 is Baidu's fifth-generation flagship language model, officially released May 9, 2026. Achieved via disaggregated fully-asynchronous reinforcement learning and scaled agentic post-training, it delivers leading performance at approximately 6% of the pre-training compute cost of comparable models — with roughly one-third the total parameters and half the active parameters of ERNIE 5.0. ERNIE 5.1 ranks #4 globally and #1 among Chinese models on the LMArena Search leaderboard (score: 1,223), with standout performance in legal reasoning, mathematics (AIME26: 99.6), and business domains. API model ID: ernie-5.1. Context: 128K tokens; max output: 65,536 tokens.

    1476

    Arena

  • Alibaba's closed-weight flagship language model, announced at the 2026 Alibaba Cloud Summit (May 20). Scored 56.6 on Artificial Analysis Intelligence Index at launch—highest-ranked Chinese model. 1M-token context with prompt caching (up to 90% discount). Pricing: $2.50/$7.50 per 1M tokens in/out.

    1475

    Arena

  • Post-training variant of GLM-5 from Z.ai (Zhipu AI) with enhanced agentic coding capabilities. Released April 7, 2026. 754B parameters (40B active) in Mixture of Experts architecture, 200K token context, 128K max output. Supports autonomous plan–execute–test–fix–optimize loops for up to 8 hours without human intervention. Trained entirely on Huawei Ascend hardware (no Nvidia). Key benchmarks: SWE-bench Pro 58.4 (world #1 at release, surpassing GPT-5.4 57.7 and Claude Opus 4.6 57.3), GPQA Diamond 86.2, AIME 2026 95.3, Terminal-Bench 2.0 63.5, MCP-Atlas 71.8, Chatbot Arena Elo 1475 (June 16, 2026, arena.ai). Available via Z.ai API ($1.40/$4.40 per 1M input/output tokens) and open weights on Hugging Face under MIT license.

    1475

    Arena

Compare Top Picks

Side-by-side comparison of the top picks by price, benchmark, and API access.
Claude Opus 4.7 vs Claude Opus 4.6Claude Opus 4.7 vs Gemini 3.1 Pro PreviewClaude Opus 4.7 vs Muse SparkClaude Opus 4.7 vs GPT-5.5Claude Opus 4.6 vs Gemini 3.1 Pro PreviewClaude Opus 4.6 vs Muse Spark

Browse Other Categories

Best LLMs for Code GenerationBest LLMs for RAGBest LLMs for Agents vs Agent Products (2026)Best LLMs for ClassificationBest Open Source LLMsBest Multimodal / Vision LLMsBest LLM for Translation in 2026Best AI Image Models in 2026Best AI Video Models in 2026Best LLMs for Reasoning & MathBest Small Language Models (SLMs)Best LLMs for JSON / Tool useCheapest LLM APIs You Can Call Right NowBest Long Context LLMsBest Mainstream LLM APIs, RankedBest LLMs for EnterpriseBest Free LLMs You Can Use Right NowBest LLMs for MarketingBest LLMs for Customer Support