mmap.page

Cursor Models

Cursor models and pricing. Prices are per million tokens. Speed, cost, and intelligence tiers come from the Cursor docs model metadata. Release year and knowledge cutoff year come from provider documentation when Cursor does not publish them. Cursor Cloud indicates availability in the Cloud Agent model picker on cursor.com/agents.

CursorBench

CursorBench 4.0 plots correctness against average cost per task. Cost decreases toward the right, so the upper-right is higher score at lower cost. Points from the same model family are connected Minimal → Low → Medium → High → Extra High → Max, skipping missing levels. An outer ring marks a model/effort available in the Cloud Agent picker. Hover or tap a point for the model name and effort. Numbers are from the public leaderboard (snapshot 4 Nov 2026). Rings on this plot are Grok 4.6/4.7, Opus 5.5, GPT-5.6 Sol, Fable 5.1, Gemini 4 Argon, Muse Wonder, Composer 2.5, and Sonnet 5.5 at all listed efforts. Gemini 3.8 Flash, Muse Spark 1.3, and DeepSeek V4.1 Flash 1017 have no rings. New arrivals: Gemini 4 Argon High 52.9% and Medium 52.8% slot between Sonnet 5.5 Extra High and Opus 5.5 Medium; Muse Wonder scales Minimal 34.2% → Max 50.4%; DeepSeek V4.1 Flash 1017 Max 47.7% at $0.98 undercuts Haiku 5.5 Max (48.4% at $1.12), and its Medium (46.3%) beats its High (45.6%). Sonnet 5.5 and Haiku 5.5 scale hard with effort: Sonnet 5.5 Low 35.8% → Max 55.5% (2nd overall); Haiku 5.5 Low 30.9% → Max 48.4%. GPT-5.6 does not: Sol Max 41.7% (tied with Muse Wonder Low) and Terra Max 41.3% sit with Haiku High (42.3%), next to Grok 4.6 Extra High and Muse Spark 1.3 Max; Luna Low is last at 16.0%.

CursorBench 4.0 score versus average cost per task Scatter plot of CursorBench 4.0 correctness against average USD cost per task. Cost decreases toward the right so the upper-right is higher score at lower cost. Points from the same model family are connected Minimal, Low, Medium, High, Extra High, then Max. An outer ring marks a model and effort available in the Cloud Agent picker. Hover or tap a point for the model name and effort. 60%55%50%45%40%35%30%25%20%15% $18$15$12$9$6$3$0 Average cost per task (higher cost on the left) CursorBench 4.0 score Opus 5.5Sonnet 5.5Gemini 4 ArgonFable 5.1Muse WonderHaiku 5.5DeepSeek V4.1 Flash 1017Opus 5Grok 4.7GLM 5.3GPT-5.6 SolMuse Spark 1.3Grok 4.6GPT-5.6 TerraGemini 3.8 FlashGLM 5.3 FlashGPT-5.6 LunaSonnet 5Composer 2.5Cursor Cloud

Model ID Context window Max context Provider Capabilities Cursor Cloud Speed Cost Intelligence Release year Knowledge cutoff year Input Cache write Cache read Output
claude-4-sonnet 200k - Anthropic Agent, Thinking, Images - medium medium high 2025 2025 $3 $3.75 $0.3 $15
claude-4-sonnet-1m - 1M Anthropic Agent, Thinking, Images - medium high high 2025 2025 $6 $7.5 $0.6 $22.5
claude-4-5-haiku 200k - Anthropic Thinking, Images - fast low moderate 2025 2025 $1 $1.25 $0.1 $5
claude-4-5-opus 200k 200k Anthropic Agent, Thinking, Images - medium high frontier 2025 2025 $5 $6.25 $0.5 $25
claude-4-5-sonnet 200k 1M Anthropic Agent, Thinking, Images - medium medium high 2025 2025 $3 $3.75 $0.3 $15
claude-4-6-opus 200k 1M Anthropic Agent, Thinking, Images - medium high frontier 2026 2025 $5 $6.25 $0.5 $25
claude-4-6-sonnet 200k 1M Anthropic Agent, Thinking, Images - medium medium high 2026 2025 $3 $3.75 $0.3 $15
claude-opus-4-7 300k 1M Anthropic Agent, Thinking, Images - medium high frontier 2026 2026 $5 $6.25 $0.5 $25
claude-fable-5 300k 1M Anthropic Agent, Thinking, Images - medium high frontier 2026 2026 $10 $12.5 $1 $50
claude-fable-5.1 300k 1M Anthropic Agent, Thinking, Images yes medium high frontier 2026 2026 $10 $12.5 $0.25 $50
claude-opus-4-7-fast 200k 1M Anthropic Agent, Thinking, Images - fast high frontier 2026 2026 $30 $37.5 $3 $150
claude-opus-4-8 300k 1M Anthropic Agent, Thinking, Images - medium high frontier 2026 2026 $5 $6.25 $0.5 $25
claude-opus-4-8-fast 300k 1M Anthropic Agent, Thinking, Images - fast high frontier 2026 2026 $10 $12.5 $1 $50
claude-opus-5 300k 1M Anthropic Agent, Thinking, Images - medium high frontier 2026 2026 $5 $6.25 $0.5 $25
claude-opus-5-fast 300k 1M Anthropic Agent, Thinking, Images - fast high frontier 2026 2026 $10 $12.5 $1 $50
claude-opus-5.5 300k 1M Anthropic Agent, Thinking, Images yes medium high frontier 2026 2026 $4 $5 $0.2 $20
claude-opus-5-5-fast 300k 1M Anthropic Agent, Thinking, Images - fast high frontier 2026 2026 $8 $10 $0.4 $40
claude-sonnet-5 200k 1M Anthropic Agent, Thinking, Images - medium medium frontier 2026 2026 $2 $2.5 $0.2 $10
claude-sonnet-5.5 200k 1M Anthropic Agent, Thinking, Images yes medium medium frontier 2026 2026 $2 $2.5 $0.1 $10
claude-haiku-5-5 300k 1M Anthropic Agent, Thinking, Images - fast low high 2026 2026 $0.1 $0.125 $0.01 $0.5
composer-1 200k - Cursor Agent, Images - medium medium high 2025 - $1.25 - $0.125 $10
composer-2.5 200k - Cursor Agent, Thinking, Images yes fast low frontier 2026 - $0.5 - $0.2 $2.5
composer-2.5-fast 200k - Cursor Agent, Thinking, Images - fast high frontier 2026 - $3 - $0.5 $15
gemini-2.5-flash 200k 1M Google Agent, Thinking, Images - fast low moderate 2025 2025 $0.3 - $0.03 $2.5
gemini-3-flash 200k 1M Google Agent, Thinking, Images - fast low moderate 2025 2025 $0.5 - $0.05 $3
gemini-3-pro 200k 1M Google Agent, Thinking, Images - medium medium high 2025 2025 $2 - $0.2 $12
gemini-3-pro-image-preview 200k 1M Google Images - medium medium high 2025 2025 $2 - $0.2 $12
gemini-3.1-pro 200k 1M Google Agent, Thinking, Images - medium medium frontier 2026 2025 $2 - $0.2 $12
gemini-3.5-flash 200k 1M Google Agent, Thinking, Images - fast low high 2026 2025 $1.5 - $0.15 $9
gemini-3.6-flash 200k 1M Google Agent, Thinking, Images - fast low high 2026 2026 $1.5 - $0.15 $7.5
gemini-3.7-flash 200k 1M Google Agent, Thinking, Images - fast low high 2026 2026 $0.75 - $0.075 $3.75
gemini-3.8-flash 200k 1M Google Agent, Thinking, Images - fast low high 2026 2026 $0.75 - $0.075 $3.75
gemini-4-argon 200k 1M Google Agent, Thinking, Images yes medium medium frontier 2026 2026 $2 - $0.1 $10
glm-5.2 200k - Z.ai Agent, Thinking - medium low high 2026 - $1.4 - $0.26 $4.4
glm-5.3 1M - Z.ai Agent, Thinking - medium low high 2026 - $1.4 - $0.26 $4.4
glm-5.3-flash 1M - Z.ai Agent, Thinking, Images - fast low moderate 2026 - $0.15 - $0.029 $0.5
gpt-5.1 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2024 $1.25 - $0.125 $10
gpt-5-fast 272k - OpenAI Agent, Thinking, Images - fast high high 2025 2024 $2.5 - $0.25 $20
gpt-5-mini 272k - OpenAI Agent, Thinking, Images - fast low moderate 2025 2024 $0.25 - $0.025 $2
gpt-5-codex 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2024 $1.25 - $0.125 $10
gpt-5.1-codex 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2024 $1.25 - $0.125 $10
gpt-5.1-codex-max 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2024 $1.25 - $0.125 $10
gpt-5.1-codex-mini 272k - OpenAI Agent, Thinking, Images - fast low moderate 2025 2024 $0.25 - $0.025 $2
gpt-5.2 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2025 $1.75 - $0.175 $14
gpt-5.2-codex 272k - OpenAI Agent, Thinking, Images - medium medium high 2025 2025 $1.75 - $0.175 $14
gpt-5.3-codex 272k - OpenAI Agent, Thinking, Images - medium medium frontier 2026 2025 $1.75 - $0.175 $14
gpt-5.4 272k 1M OpenAI Agent, Thinking, Images - medium medium frontier 2026 2025 $2.5 - $0.25 $15
gpt-5.4-fast 272k 1M OpenAI Agent, Thinking, Images - fast high frontier 2026 2025 $5 - $0.5 $30
gpt-5.4-mini 272k - OpenAI Agent, Thinking, Images - fast low high 2026 2025 $0.75 - $0.075 $4.5
gpt-5.4-nano 272k - OpenAI Agent, Thinking, Images - fast low moderate 2026 2025 $0.2 - $0.02 $1.25
gpt-5.5 272k 1M OpenAI Agent, Thinking, Images - medium medium frontier 2026 2025 $5 - $0.5 $30
gpt-5.5-fast 272k 1M OpenAI Agent, Thinking, Images - fast high frontier 2026 2025 $12.5 - $1.25 $75
gpt-5.6-luna 272k 1M OpenAI Agent, Thinking, Images - fast low moderate 2026 2026 $0.2 $0.25 $0.02 $1.2
gpt-5.6-luna-fast 272k 1M OpenAI Agent, Thinking, Images - fast low moderate 2026 2026 $0.4 $0.5 $0.04 $2.4
gpt-5.6-sol 272k 1M OpenAI Agent, Thinking, Images yes medium medium frontier 2026 2026 $4 $5 $0.4 $20
gpt-5.6-sol-fast 272k 1M OpenAI Agent, Thinking, Images yes fast high frontier 2026 2026 $8 $10 $0.8 $40
gpt-5.6-terra 272k 1M OpenAI Agent, Thinking, Images - fast medium high 2026 2026 $2 $2.5 $0.2 $12
gpt-5.6-terra-fast 272k 1M OpenAI Agent, Thinking, Images - fast high high 2026 2026 $4 $5 $0.4 $24
grok-4.5 256k - Cursor Agent, Thinking - medium medium frontier 2026 2026 $2 - $0.5 $6
grok-4.5-fast 256k - Cursor Agent, Thinking - fast high frontier 2026 2026 $4 - $1 $18
grok-4.6 256k - Cursor Agent, Thinking yes medium medium frontier 2026 2026 $2 - $0.5 $6
grok-4.6-fast 256k - Cursor Agent, Thinking yes fast high frontier 2026 2026 $4 - $1 $12
grok-4.7 256k 500k Cursor Agent, Thinking yes medium medium frontier 2026 2026 $2 - $0.5 $6
grok-4.7-fast 256k 500k Cursor Agent, Thinking yes fast high frontier 2026 2026 $4 - $1 $12
muse-spark-1.3 300k 1M Meta Agent, Thinking, Images - medium low frontier 2026 - $1.25 - $0.15 $4.25
muse-wonder-1.0 300k 1M Meta Agent, Thinking, Images yes medium high frontier 2026 - $4 - $0.2 $20
kimi-k2.7-code 262k - Moonshot Agent, Thinking, Images - medium low high 2026 - $0.95 - $0.19 $4
kimi-k3 200k 1M Moonshot Agent, Thinking, Images - medium medium high 2026 - $3 - $0.3 $15
deepseek-v4.1-flash-1017 1M - DeepSeek Agent, Thinking - fast medium high 2026 - $0.3 - $0.006 $1.2

Haiku 5.5 prices in the table are the rates for requests up to 100k input tokens. Above 100k, Cursor bills 5x ($0.5 input, $0.625 cache write, $0.05 cache read, $2.5 output), which applies inside the 300k default window and at the 1M max.