All promotions
GuidesMCP

Claude Haiku 5.5

AnthropicReleased Oct 7, 2026Closed weightsclaude-haiku-5-5

Intelligence (Epoch)
159 · 13th of 111
Knowledge cutoff
Jun 2026
Context
1M tokens
Max output
128K tokens
Input
text, image
Anthropic's API
yes, as claude-haiku-5-5

Read Oct 7, 00:00 UTC from the hosts' published offers and OpenRouter, spec sheet from Anthropic · cutoff: platform.claude.com

Input, per M$0.10at Anthropic
Cache, per M$0.01at Anthropic
Providers1its maker's own price, for reference

Intelligence for the price

On the best-deal line: no model scores higher for less than $0.60 per M.

100120140160180$0.10$1$10$100GPT-4o mini: 126 at $0.75gpt-oss-20b: 138 at $0.12GPT-4.1 mini: 135 at $2.00Gemini 3.6 Flash: 155 at $4.50Llama 3.3 70B: 127 at $0.42Llama 3.1 8B: 116 at $0.06gpt-oss-120b: 140 at $0.18DeepSeek Chat: 146 at $1.21DeepSeek R1: 139 at $2.65Llama 4 Scout: 130 at $0.40Llama 4 Maverick: 132 at $1.00Qwen 2.5 72B: 129 at $0.76Gemma 3 27B: 130 at $0.24DeepSeek V3.2: 146 at $0.52Kimi K2.6: 151 at $2.40GLM 5.2: 152 at $2.36Qwen3.8 27B: 152 at $1.29Gemini 3.8 Flash: 157 at $4.50Gemini 3.7 Flash: 158 at $4.50Gemini 3.5 Flash-Lite: 145 at $2.80DeepSeek V4 Pro 0813: 156 at $2.50DeepSeek V4.1 Flash: 156 at $0.47GLM-5.3: 156 at $2.36Grok 4.6: 157 at $8.00Grok 4.7: 154 at $8.00Grok 4.5: 154 at $8.00Claude Opus 5: 164 at $30.00Claude Opus 5.5: 169 at $24.00Claude Sonnet 5: 157 at $12.00Claude Fable 5: 163 at $60.00Claude Fable 5.1: 166 at $60.00GPT-6 Astra: 169 at $60.00GPT-6 Sol: 164 at $12.00GPT-6 Luna: 157 at $0.60GPT-5.6 Sol: 163 at $24.00GPT-5.6 Terra: 160 at $14.00GPT-5.6 Luna: 156 at $1.40DeepSeek V4 Flash 0731: 155 at $0.21GLM-5.3-Flash: 153 at $0.19Kimi K3: 158 at $11.35DeepSeek V4 Pro 0423: 149 at $2.40Kimi K2.7 Code: 150 at $3.71DeepSeek V4 Flash 0423: 146 at $0.21GLM 5.1: 150 at $4.00Gemma 4 31B: 143 at $0.38MiniMax-M3: 147 at $1.19Qwen3 235B A22B Instruct 2507: 139 at $0.44Qwen3.6 35B A3B: 144 at $1.00MiniMax M2.7: 146 at $1.05Qwen3.5 397B A17B: 147 at $3.45GLM 5: 146 at $2.52Gemma 4 26B A4B: 142 at $0.26MiniMax M2.5: 147 at $1.22Qwen3.5-35B-A3B: 142 at $1.14Kimi K2.5: 148 at $2.70DeepSeek V3.1: 140 at $1.20Qwen3.6 27B: 146 at $2.25GLM 4.7: 144 at $2.15Qwen3.5 9B: 139 at $0.25Qwen3 30B A3B: 136 at $0.24Nemotron 3 Ultra: 146 at $2.70Mistral Nemo: 119 at $0.038Mistral Medium 3.5: 141 at $9.00DeepSeek V3.2 Exp: 145 at $0.34Qwen3 235B A22B Thinking 2507: 144 at $2.53Mistral Medium 3: 134 at $2.40Qwen3 14B: 138 at $0.32Llama 3.1 70B Instruct: 126 at $0.80Mixtral 8x22B Instruct: 122 at $8.00Kimi K2 Thinking: 146 at $3.10Qwen3 32B: 138 at $0.36DeepSeek V3: 132 at $1.21Inkling: 148 at $5.00Inkling Small: 150 at $1.65Qwen3 Max: 142 at $7.20Gemma 3 12B: 124 at $0.15Qwen3.8 Max: 157 at $8.00Qwen3.7 Max: 154 at $5.00Qwen3.7 Flash: 144 at $0.16Qwen3.7 Plus: 147 at $2.00Qwen3.5 Plus 2026-04-20: 146 at $2.80Qwen3.6 Flash: 143 at $1.75Qwen3.6 Max Preview: 149 at $9.10Qwen3.6 Plus: 147 at $3.50Qwen3.5-Flash: 144 at $0.50Qwen3 30B A3B Thinking 2507: 140 at $2.60Qwen3 8B: 136 at $0.88Mistral Small 3.1 24B: 128 at $0.91Gemma 3 4B: 116 at $0.09Phi 4: 130 at $0.21Qwen2.5 Coder 32B Instruct: 119 at $1.66Qwen2.5 7B Instruct: 119 at $0.30Command R+ (08-2024): 119 at $12.50Gemma 2 27B: 122 at $1.30Claude Sonnet 5.5: 166 at $12.00GPT-6.1 Sol: 168 at $12.00Gemini 4 Argon: 166 at $12.00Mistral Large 4: 150 at $5.54Muse Wonder: 162 at $24.00Muse Spark 1.3: 157 at $5.50Muse Spark Open: 155 at $5.50GPT-4.1: 137 at $10.00GPT-5.3 Codex: 157 at $15.75Claude Opus 4.1: 144 at $90.00Claude Sonnet 4.5: 147 at $18.00Claude Haiku 4.5: 142 at $6.00Gemini 2.5 Pro: 145 at $11.25Gemini 2.5 Flash: 141 at $2.80Gemini 3.1 Pro Preview: 155 at $14.00Gemini 3 Flash Preview: 152 at $3.50Claude Haiku 5.5 · 159

Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 7. Score: Epoch AI, CC BY 4.0, 159 with an interval of 156 to 162.

Against its rivals

Its previous version, the best of its price range and the top score

Same level for 2.9× less: DeepSeek V4 Flash 0731 at Relace, $0.21 per M input + output · their Epoch intervals overlap (155 against 159).

Bars: Epoch Capabilities Index (Epoch AI, CC BY 4.0), drawn from 100, below the lowest score in the catalogue · price: lowest input + output per M.

ProviderInputOutputCache ↑Checked
Anthropic$0.10$0.50$0.01Oct 7

1 provider · prices per million tokens, read Oct 7, 00:00 UTC

ProviderWeightsContextMax outputToolsStructured outputsReasoning
Anthropic—1M125K———

As served by each host and declared to OpenRouter; not verified by us · weights: the precision the host declares

Claude Haiku 5.5 spec sheet, field by field, with sources

Claude Haiku 5.5 spec sheetmaker-declared

What Anthropic declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.

Context window
1,000,000 tokens
Parameters
not published by the maker
Active experts
not published by the maker
Weight precision
not applicable — no weights publishedcurated
License
Proprietary — API onlycuratedScope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
Knowledge cutoff
June 2026
Modalities
text · image → text
Max output (official)
128,000 tokens
Release date
October 7, 2026
Maker lifecycle status
Active (latest)

platform.claude.com · maker's documentation · read 2026-10-07 · 7 of 10 fields published by the maker

ProviderChatbot 1,000 in / 500 outRAG / long context 8,000 in / 500 outBatch summarize 4,000 in / 1,000 out
Anthropic$0.35$1.05$0.90

What each host bills for 1,000 requests of a representative job

Your own workload

    Other Anthropic models · all its models, hosts and history ›