/best/writingupdated May 202685 models evaluated
Best AI for writing
Long-form drafting, rewriting, tone-matching, and editorial work.
Writing is where models diverge most on taste. Two models can score identically on instruction-following and produce prose with completely different voice — one tight and readable, the other syrupy and evasive. Benchmark scores alone won't tell you which model matches your brand.
What matters: rhetorical range, willingness to cut, instruction adherence at the paragraph level, and refusal discipline. Good models don't moralize at the drop of a hat. We weight side-by-side human preference strongly here, along with tone consistency across long documents.
Our writing rank weights writing quality (2.5×), instruction-following (1.5×), and reasoning (0.5×). For editorial work at scale, instruction-following is the main differentiator — models that silently ignore style constraints waste editorial time.
Full rankings
All 85 models, scored for writing
| # | Model | Provider | Task score | $/in | $/out | Context |
|---|---|---|---|---|---|---|
| 01 | DeepSeek V4 Flash | DDeepSeek | 4.50/5.0 | $0.100 | $0.200 | 1.0M |
| 02 | GLM-4.7 Flash | ZZhipu AI | 4.50/5.0 | $0.060 | $0.400 | 203K |
| 03 | NVIDIA: Nemotron 3 Super | NNVIDIA | 4.50/5.0 | $0.090 | $0.450 | 1M |
| 04 | DeepSeek V4 Pro | DDeepSeek | 4.50/5.0 | $0.435 | $0.870 | 1.0M |
| 05 | Qwen: Qwen3.6 Plus | QQwen | 4.50/5.0 | $0.325 | $1.95 | 1M |
| 06 | Qwen: Qwen3.6 27B | QQwen | 4.50/5.0 | $0.300 | $2.00 | 262K |
| 07 | Devstral 2 2512 | MMistral | 4.50/5.0 | $0.400 | $2.00 | 262K |
| 08 | MiMo-V2.5 | XXiaomi | 4.50/5.0 | $0.400 | $2.00 | 1.0M |
| 09 | R1 0528 | DDeepSeek | 4.50/5.0 | $0.500 | $2.15 | 164K |
| 10 | R1 | DDeepSeek | 4.50/5.0 | $0.700 | $2.50 | 164K |
| 11 | Grok 4.3 | XXAI | 4.50/5.0 | $1.25 | $2.50 | 1M |
| 12 | Gemini 3 Flash Preview | GGoogle | 4.50/5.0 | $0.500 | $3.00 | 1.0M |
| 13 | MoonshotAI: Kimi K2.6 | MMoonshotAI | 4.50/5.0 | $0.730 | $3.49 | 262K |
| 14 | Qwen: Qwen3.6 Max Preview | QQwen | 4.50/5.0 | $1.04 | $6.24 | 262K |
| 15 | Qwen 3.7 Max | QQwen | 4.50/5.0 | $2.50 | $7.50 | 1M |
| 16 | Gemini 3.5 Flash | GGoogle | 4.50/5.0 | $1.50 | $9.00 | 1.0M |
| 17 | Gemini 3.1 Pro Preview | GGoogle | 4.50/5.0 | $2.00 | $12.00 | 1.0M |
| 18 | GPT-5.2 | OOpenAI | 4.50/5.0 | $1.75 | $14.00 | 400K |
| 19 | Claude Opus 4.7 | AAnthropic | 4.50/5.0 | $5.00 | $25.00 | 1M |
| 20 | GPT-5.5 | OOpenAI | 4.50/5.0 | $5.00 | $30.00 | 1.1M |
| 21 | Qwen: Qwen3 235B A22B Instruct 2507 | QQwen | 4.00/5.0 | $0.071 | $0.100 | 262K |
| 22 | Ministral 3 3B 2512 | MMistral | 4.00/5.0 | $0.100 | $0.100 | 131K |
| 23 | Qwen: Qwen3.5-9B | QQwen | 4.00/5.0 | $0.040 | $0.150 | 262K |
| 24 | OpenAI: gpt-oss-120b | OOpenAI | 4.00/5.0 | $0.039 | $0.180 | 131K |
| 25 | Ministral 3 8B 2512 | MMistral | 4.00/5.0 | $0.150 | $0.150 | 262K |
| 26 | Ministral 3 14B 2512 | MMistral | 4.00/5.0 | $0.200 | $0.200 | 262K |
| 27 | Qwen: Qwen3 30B A3B Instruct 2507 | QQwen | 4.00/5.0 | $0.090 | $0.300 | 262K |
| 28 | Gemma 4 31B | GGoogle | 4.00/5.0 | $0.120 | $0.370 | 262K |
| 29 | DeepSeek V3.2 | DDeepSeek | 4.00/5.0 | $0.252 | $0.378 | 131K |
| 30 | Grok 4.1 Fast | XXAI | 4.00/5.0 | $0.200 | $0.500 | 2M |
| 31 | DeepSeek V3.1 | DDeepSeek | 4.00/5.0 | $0.210 | $0.790 | 164K |
| 32 | Qwen: Qwen3.5-35B-A3B | QQwen | 4.00/5.0 | $0.139 | $1.00 | 262K |
| 33 | Qwen: Qwen3.6 35B A3B | QQwen | 4.00/5.0 | $0.150 | $1.00 | 262K |
| 34 | Qwen: Qwen3.6 Flash | QQwen | 4.00/5.0 | $0.188 | $1.13 | 1M |
| 35 | GPT-5.4 Nano | OOpenAI | 4.00/5.0 | $0.200 | $1.25 | 400K |
| 36 | Gemini 3.1 Flash Lite Preview | GGoogle | 4.00/5.0 | $0.250 | $1.50 | 1.0M |
| 37 | Google: Gemini 3.1 Flash Lite | GGoogle | 4.00/5.0 | $0.250 | $1.50 | 1.0M |
| 38 | Qwen: Qwen3.5 Plus 2026-04-20 | QQwen | 4.00/5.0 | $0.300 | $1.80 | 1M |
| 39 | GPT-5 Mini | OOpenAI | 4.00/5.0 | $0.250 | $2.00 | 400K |
| 40 | Mistral Medium 3.1 | MMistral | 4.00/5.0 | $0.400 | $2.00 | 131K |
| 41 | Xiaomi: MiMo-V2-Omni | XXiaomi | 4.00/5.0 | $0.400 | $2.00 | 262K |
| 42 | xAI: Grok Build 0.1 | XXAI | 4.00/5.0 | $1.00 | $2.00 | 256K |
| 43 | Gemini 2.5 Flash | GGoogle | 4.00/5.0 | $0.300 | $2.50 | 1.0M |
| 44 | Grok 4.20 | XXAI | 4.00/5.0 | $1.25 | $2.50 | 2M |
| 45 | MiMo-V2.5-Pro | XXiaomi | 4.00/5.0 | $1.00 | $3.00 | 1.0M |
| 46 | GPT-5.4 Mini | OOpenAI | 4.00/5.0 | $0.750 | $4.50 | 400K |
| 47 | o4 Mini | OOpenAI | 4.00/5.0 | $1.10 | $4.40 | 200K |
| 48 | Mistral Medium 3.5 | MMistral | 4.00/5.0 | $1.50 | $7.50 | 262K |
| 49 | GPT-4.1 | OOpenAI | 4.00/5.0 | $2.00 | $8.00 | 1.0M |
| 50 | o3 | OOpenAI | 4.00/5.0 | $2.00 | $8.00 | 200K |
| 51 | GPT-5 | OOpenAI | 4.00/5.0 | $1.25 | $10.00 | 400K |
| 52 | Gemini 2.5 Pro | GGoogle | 4.00/5.0 | $1.25 | $10.00 | 1.0M |
| 53 | GPT-5.1 | OOpenAI | 4.00/5.0 | $1.25 | $10.00 | 400K |
| 54 | GPT-5.4 | OOpenAI | 4.00/5.0 | $2.50 | $15.00 | 1.1M |
| 55 | Claude Sonnet 4.6 | AAnthropic | 4.00/5.0 | $3.00 | $15.00 | 1M |
| 56 | Claude Opus 4.6 | AAnthropic | 4.00/5.0 | $5.00 | $25.00 | 1M |
| 57 | Xiaomi: MiMo-V2-Flash | XXiaomi | 3.50/5.0 | $0.100 | $0.300 | 262K |
| 58 | Gemma 4 26B A4B | GGoogle | 3.50/5.0 | $0.060 | $0.330 | 262K |
| 59 | Gemini 2.5 Flash Lite | GGoogle | 3.50/5.0 | $0.100 | $0.400 | 1.0M |
| 60 | Grok 3 Mini | XXAI | 3.50/5.0 | $0.300 | $0.500 | 131K |
| 61 | Mistral Small 4 | MMistral | 3.50/5.0 | $0.150 | $0.600 | 262K |
| 62 | DeepSeek V3.1 Terminus | DDeepSeek | 3.50/5.0 | $0.270 | $0.950 | 164K |
| 63 | GPT-4.1 Mini | OOpenAI | 3.50/5.0 | $0.400 | $1.60 | 1.0M |
| 64 | GLM-4.7 | ZZhipu AI | 3.50/5.0 | $0.400 | $1.75 | 203K |
| 65 | Xiaomi: MiMo-V2-Pro | XXiaomi | 3.50/5.0 | $1.00 | $3.00 | 1.0M |
| 66 | Claude Haiku 4.5 | AAnthropic | 3.50/5.0 | $1.00 | $5.00 | 200K |
| 67 | Grok 4 | XXAI | 3.50/5.0 | $3.00 | $15.00 | 256K |
| 68 | OpenAI: gpt-oss-20b | OOpenAI | 3.00/5.0 | $0.030 | $0.140 | 131K |
| 69 | Mistral Small 3.2 24B | MMistral | 3.00/5.0 | $0.075 | $0.200 | 128K |
| 70 | Llama 4 Scout | MMeta | 3.00/5.0 | $0.080 | $0.300 | 10M |
| 71 | Llama 3.3 70B Instruct | MMeta | 3.00/5.0 | $0.100 | $0.320 | 131K |
| 72 | GPT-5 Nano | OOpenAI | 3.00/5.0 | $0.050 | $0.400 | 400K |
| 73 | GPT-4.1 Nano | OOpenAI | 3.00/5.0 | $0.100 | $0.400 | 1.0M |
| 74 | Llama 4 Maverick | MMeta | 3.00/5.0 | $0.150 | $0.600 | 1.0M |
| 75 | Grok Code Fast 1 | XXAI | 3.00/5.0 | $0.200 | $1.50 | 256K |
| 76 | Mistral Large 3 2512 | MMistral | 3.00/5.0 | $0.500 | $1.50 | 262K |
| 77 | GPT-4o | OOpenAI | 3.00/5.0 | $2.50 | $10.00 | 128K |
| 78 | Grok 3 | XXAI | 3.00/5.0 | $3.00 | $15.00 | 131K |
| 79 | NVIDIA: Nemotron 3 Nano 30B A3B | NNVIDIA | 2.50/5.0 | $0.050 | $0.200 | 262K |
| 80 | Qwen: Qwen3 Coder 30B A3B Instruct | QQwen | 2.50/5.0 | $0.070 | $0.270 | 160K |
| 81 | Devstral Small 1.1 | MMistral | 2.50/5.0 | $0.100 | $0.300 | 131K |
| 82 | GPT-4o-mini | OOpenAI | 2.50/5.0 | $0.150 | $0.600 | 128K |
| 83 | Mistral Small 3.1 24B | MMistral | 2.50/5.0 | $0.351 | $0.555 | 128K |
| 84 | Codestral 2508 | MMistral | 2.50/5.0 | $0.300 | $0.900 | 256K |
| 85 | Devstral Medium | MMistral | 2.50/5.0 | $0.400 | $2.00 | 131K |
Pricing — top 5 for writing
D
4.50/5.0
Z
4.50/5.0
N
4.50/5.0
D
4.50/5.0
Q
4.50/5.0
The best AI for writing changes every month.
We'll email you when rankings shift, new models hit the top 5, or pricing cuts reshuffle the value leaders.
Get notified when models change
Price drops, new models, benchmark updates. One email per change, no spam.