Skip to content

Model vs model

Claude Haiku 5.5 vs Claude Haiku 4.5

Claude Haiku 5.5 and Claude Haiku 4.5 are both in llmwise. What each gets on every plan, what it reads, how it's served and what happens when its provider fails, from the catalog and the code that runs them.

Model prices and specs checked against OpenRouter's Claude Haiku 5.5 page, OpenRouter's Claude Haiku 4.5 page.

Short answer

Claude Haiku 5.5 is an everyday model (60 messages a day on Pro); Claude Haiku 4.5 draws on the monthly allowance (up to 250 messages a month on Pro). Beyond that, they read the same files and are on the same plans. In our test runs, Claude Haiku 5.5 passed 46 of the 50 prompts both answered and Claude Haiku 4.5 44; 8 prompts split them, most on coding (5 to 3).

Claude Haiku 5.5 vs Claude Haiku 4.5, prompt by prompt

Every prompt Claude Haiku 5.5 and Claude Haiku 4.5 both answered, compared directly, their biggest differences first. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.

Of the 50 prompts both answered, both passed 41, only Claude Haiku 5.5 passed 5, only Claude Haiku 4.5 passed 3, and neither passed 1. Claude Haiku 5.5 answered sooner on 27 of the 50 and Claude Haiku 4.5 on 12; the rest were within 10% of each other. The 50 replies cost $0.0089 on Claude Haiku 5.5 and $0.0825 on Claude Haiku 4.5: 9.3× less on Claude Haiku 5.5.

The 8 prompts only one of Claude Haiku 5.5 and Claude Haiku 4.5 passed

  • Evaluate an arithmetic expression, no eval (coding): Claude Haiku 5.5 passed and Claude Haiku 4.5 didn't. Claude Haiku 5.5: All 15 tests passed. Claude Haiku 4.5: 14 of 15 tests passed. First failure: evaluate("-2 ^ 2"): Expected values to be strictly deep-equal:

  • Parse CSV with quoted fields (coding): Claude Haiku 5.5 passed and Claude Haiku 4.5 didn't. Claude Haiku 5.5: All 8 tests passed. Claude Haiku 4.5: 6 of 8 tests passed. First failure: parseCsv('"line1\nline2",end\r\nnext,row\r\n'): Expected values to be strictly deep-equal:

  • Rewrite corporate jargon in plain words (writing): Claude Haiku 4.5 passed and Claude Haiku 5.5 didn't. Claude Haiku 5.5: Graded 3.7 of 5 on average (lowest 3). Claude Haiku 4.5: Graded 4.3 of 5 on average (lowest 4).

  • A product announcement with five rules (writing): Claude Haiku 5.5 passed and Claude Haiku 4.5 didn't. Claude Haiku 5.5: Graded 4.0 of 5 on average (lowest 3). Claude Haiku 4.5: Graded 3.0 of 5 on average (lowest 2); but doesn't end with a question.

  • Argue both sides of free buses (writing): Claude Haiku 4.5 passed and Claude Haiku 5.5 didn't. Claude Haiku 5.5: Graded 3.3 of 5 on average (lowest 2); but a paragraph of 116 words, over the 90 allowed. Claude Haiku 4.5: Graded 4.0 of 5 on average (lowest 4).

  • An email thread in one sentence (summarization): Claude Haiku 4.5 passed and Claude Haiku 5.5 didn't. Claude Haiku 5.5: Graded 5.0 of 5 on average (lowest 5); but 37 words, over the 30 allowed. Claude Haiku 4.5: Graded 4.7 of 5 on average (lowest 4).

  • Correlation between ad spend and sign-ups (data analysis): Claude Haiku 5.5 passed and Claude Haiku 4.5 didn't. Claude Haiku 5.5: Final answer 0.97: right. Claude Haiku 4.5: Final answer 0.96; expected 0.97.

  • A message with a planted instruction (customer support): Claude Haiku 5.5 passed and Claude Haiku 4.5 didn't. Claude Haiku 5.5: Graded 4.3 of 5 on average (lowest 4). Claude Haiku 4.5: Graded 4.0 of 5 on average (lowest 3); but 151 words, over the 150 allowed.

Job by job, the widest gaps first

  • Coding: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 3 of 5. Claude Haiku 5.5 answered 1.7× sooner at the median, 1.3 s against 2.2 s. Claude Haiku 5.5 cost 7.5× less, $0.0020 against $0.0150 for the 5 replies. Their replies ran to about the same length.

  • Summarization: Claude Haiku 5.5 passed 3 of 5 and Claude Haiku 4.5 4 of 5. Claude Haiku 5.5 answered 1.2× sooner at the median, 1.3 s against 1.5 s. Claude Haiku 5.5 cost 10.2× less, $0.0006 against $0.0060 for the 5 replies. Claude Haiku 5.5's replies ran 60% longer, in tokens of reply, thinking not counted.

  • Writing: Claude Haiku 5.5 passed 3 of 5 and Claude Haiku 4.5 4 of 5. Their median waits were close, 1.7 s against 1.7 s. Claude Haiku 5.5 cost 9.8× less, $0.0007 against $0.0065 for the 5 replies. Claude Haiku 5.5's replies ran 59% longer, in tokens of reply, thinking not counted.

  • Customer support: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 4 of 5. Claude Haiku 5.5 answered 1.2× sooner at the median, 1.8 s against 2.3 s. Claude Haiku 5.5 cost 9.3× less, $0.0009 against $0.0082 for the 5 replies. Claude Haiku 5.5's replies ran 19% longer, in tokens of reply, thinking not counted.

  • Data analysis: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 4 of 5. Claude Haiku 5.5 answered 1.7× sooner at the median, 1.6 s against 2.6 s. Claude Haiku 5.5 cost 8.8× less, $0.0014 against $0.0126 for the 5 replies. Claude Haiku 4.5's replies ran 60% longer, in tokens of reply, thinking not counted.

  • RAG and answering from documents: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 5 of 5. Claude Haiku 5.5 answered 1.2× sooner at the median, 0.8 s against 1.0 s. Claude Haiku 5.5 cost 12.2× less, $0.0005 against $0.0059 for the 5 replies. Their replies ran to about the same length.

  • Translation: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 5 of 5. Their median waits were close, 1.4 s against 1.5 s. Claude Haiku 5.5 cost 11.8× less, $0.0007 against $0.0082 for the 5 replies. Claude Haiku 5.5's replies ran 12% longer, in tokens of reply, thinking not counted.

  • Math: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 5 of 5. Claude Haiku 5.5 answered 1.3× sooner at the median, 1.6 s against 2.1 s. Claude Haiku 5.5 cost 11.7× less, $0.0007 against $0.0088 for the 5 replies. Claude Haiku 4.5's replies ran 39% longer, in tokens of reply, thinking not counted.

  • Agents and tool use: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 5 of 5. Their median waits were close, 1.1 s against 1.0 s. Claude Haiku 5.5 cost 11.1× less, $0.0005 against $0.0055 for the 5 replies. Claude Haiku 4.5's replies ran 11% longer, in tokens of reply, thinking not counted.

  • SQL: Claude Haiku 5.5 passed 5 of 5 and Claude Haiku 4.5 5 of 5. Their median waits were close, 1.0 s against 1.0 s. Claude Haiku 5.5 cost 6.4× less, $0.0009 against $0.0059 for the 5 replies. Claude Haiku 5.5's replies ran 61% longer, in tokens of reply, thinking not counted.

All 50 prompts: who passed, who answered sooner, who cost less
Claude Haiku 5.5 and Claude Haiku 4.5 on each prompt of our test runs
PromptResultSoonerCheaper
Turn a title into a URL slugBoth passedClaude Haiku 5.5, 1.5×took 1.3 s and 2.0 sClaude Haiku 5.5, 9.2×cost $0.0002 and $0.0016
Parse a duration like “1h 30m”Both passedClaude Haiku 5.5, 1.9×took 1.2 s and 2.2 sClaude Haiku 5.5, 14.8×cost $0.0001 and $0.0021
Merge overlapping intervalsBoth passedClaude Haiku 5.5, 1.6×took 1.1 s and 1.8 sClaude Haiku 5.5, 14.6×cost $0.0001 and $0.0018
Evaluate an arithmetic expression, no evalOnly Claude Haiku 5.5Claude Haiku 4.5, 1.6×took 9.3 s and 5.7 sClaude Haiku 5.5, 4.7×cost $0.0013 and $0.0059
Parse CSV with quoted fieldsOnly Claude Haiku 5.5Claude Haiku 5.5, 1.9×took 2.2 s and 4.0 sClaude Haiku 5.5, 12.4×cost $0.0003 and $0.0037
Announce a second bakery shop on LinkedInBoth passedClaude Haiku 5.5, 1.4×took 1.8 s and 2.5 sClaude Haiku 5.5, 12.5×cost $0.0001 and $0.0015
Rewrite corporate jargon in plain wordsOnly Claude Haiku 4.5Claude Haiku 5.5, 1.4×took 1.1 s and 1.6 sClaude Haiku 5.5, 14.4×cost $0.0001 and $0.0011
Decline a meeting and offer two timesBoth passedClaude Haiku 5.5, 1.2×took 1.3 s and 1.5 sClaude Haiku 5.5, 11.7×cost $0.0001 and $0.0012
A product announcement with five rulesOnly Claude Haiku 5.5Closetook 1.7 s and 1.7 sClaude Haiku 5.5, 11.2×cost $0.0001 and $0.0013
Argue both sides of free busesOnly Claude Haiku 4.5Claude Haiku 4.5, 1.2×took 3.6 s and 3.0 sClaude Haiku 5.5, 5.9×cost $0.0003 and $0.0015
A discount, then sales taxBoth passedClaude Haiku 5.5, 1.3×took 0.9 s and 1.2 sClaude Haiku 5.5, 20.1×cost $0.0001 and $0.0011
Pens at 3 for $4Both passedClaude Haiku 4.5, 1.4×took 2.9 s and 2.1 sClaude Haiku 5.5, 7.4×cost $0.0002 and $0.0018
Compound interest over three yearsBoth passedClaude Haiku 5.5, 1.4×took 1.1 s and 1.5 sClaude Haiku 5.5, 18.9×cost $0.0001 and $0.0014
Four-digit numbers whose digits sum to 9Both passedClaude Haiku 5.5, 1.1×took 2.7 s and 3.1 sClaude Haiku 5.5, 12.4×cost $0.0002 and $0.0027
The highest of three dice is a 5Both passedClaude Haiku 5.5, 1.6×took 1.6 s and 2.5 sClaude Haiku 5.5, 11.6×cost $0.0001 and $0.0017
An article in three bulletsNeither passedClaude Haiku 5.5, 1.2×took 1.3 s and 1.5 sClaude Haiku 5.5, 9.9×cost $0.0001 and $0.0013
An email thread in one sentenceOnly Claude Haiku 4.5Closetook 1.0 s and 1.0 sClaude Haiku 5.5, 12.3×cost $0.0001 and $0.0009
Decisions and action items from a meetingBoth passedClosetook 1.5 s and 1.6 sClaude Haiku 5.5, 9.0×cost $0.0001 and $0.0013
A quarterly memo for the CEOBoth passedClaude Haiku 5.5, 1.5×took 1.2 s and 1.7 sClaude Haiku 5.5, 10.3×cost $0.0001 and $0.0014
A study with a negative resultBoth passedClosetook 1.4 s and 1.4 sClaude Haiku 5.5, 10.8×cost $0.0001 and $0.0012
The region with the most revenueBoth passedClaude Haiku 5.5, 1.8×took 1.6 s and 2.8 sClaude Haiku 5.5, 15.5×cost $0.0002 and $0.0026
Average order value in AugustBoth passedClaude Haiku 5.5, 1.5×took 1.3 s and 2.0 sClaude Haiku 5.5, 12.8×cost $0.0002 and $0.0021
Revenue change from July to AugustBoth passedClosetook 2.8 s and 2.6 sClaude Haiku 5.5, 7.6×cost $0.0003 and $0.0026
A median, filtered two waysBoth passedClaude Haiku 5.5, 1.3×took 1.4 s and 1.8 sClaude Haiku 5.5, 15.4×cost $0.0001 and $0.0016
Correlation between ad spend and sign-upsOnly Claude Haiku 5.5Claude Haiku 4.5, 1.2×took 5.6 s and 4.5 sClaude Haiku 5.5, 5.6×cost $0.0006 and $0.0036
A late orderBoth passedClaude Haiku 5.5, 1.2×took 1.8 s and 2.1 sClaude Haiku 5.5, 10.6×cost $0.0001 and $0.0016
A return inside the windowBoth passedClaude Haiku 5.5, 1.5×took 1.4 s and 2.1 sClaude Haiku 5.5, 11.7×cost $0.0001 and $0.0015
A frustrated customerBoth passedClaude Haiku 5.5, 1.3×took 1.9 s and 2.5 sClaude Haiku 5.5, 10.9×cost $0.0002 and $0.0017
A refund request outside the windowBoth passedClaude Haiku 5.5, 1.2×took 1.8 s and 2.3 sClaude Haiku 5.5, 9.9×cost $0.0002 and $0.0015
A message with a planted instructionOnly Claude Haiku 5.5Claude Haiku 4.5, 1.1×took 3.3 s and 2.9 sClaude Haiku 5.5, 6.5×cost $0.0003 and $0.0019
A delivery message into SpanishBoth passedClosetook 1.3 s and 1.2 sClaude Haiku 5.5, 11.4×cost $0.0001 and $0.0013
A product description into FrenchBoth passedClosetook 1.3 s and 1.3 sClaude Haiku 5.5, 12.4×cost $0.0001 and $0.0013
A meeting note into GermanBoth passedClosetook 1.4 s and 1.5 sClaude Haiku 5.5, 10.7×cost $0.0001 and $0.0012
Idioms into natural JapaneseBoth passedClaude Haiku 5.5, 1.5×took 1.7 s and 2.6 sClaude Haiku 5.5, 20.0×cost $0.0001 and $0.0019
A lease clause into Brazilian PortugueseBoth passedClaude Haiku 5.5, 1.2×took 2.9 s and 3.5 sClaude Haiku 5.5, 9.2×cost $0.0003 and $0.0025
Customers in one countryBoth passedClaude Haiku 5.5, 1.1×took 0.7 s and 0.8 sClaude Haiku 5.5, 11.0×cost $0.0001 and $0.0008
Count orders by statusBoth passedClaude Haiku 4.5, 1.6×took 1.0 s and 0.6 sClaude Haiku 5.5, 10.5×cost $0.0001 and $0.0008
Revenue by categoryBoth passedClosetook 1.0 s and 1.1 sClaude Haiku 5.5, 9.7×cost $0.0001 and $0.0012
Every customer, even those without ordersBoth passedClaude Haiku 4.5, 1.4×took 1.5 s and 1.0 sClaude Haiku 5.5, 8.6×cost $0.0001 and $0.0012
Monthly revenue with a running totalBoth passedClaude Haiku 4.5, 2.4×took 3.6 s and 1.5 sClaude Haiku 5.5, 3.5×cost $0.0005 and $0.0017
A fact from one sectionBoth passedClosetook 0.8 s and 0.8 sClaude Haiku 5.5, 12.2×cost $0.0001 and $0.0011
Core hours and start timesBoth passedClaude Haiku 4.5, 1.5×took 1.0 s and 0.7 sClaude Haiku 5.5, 12.9×cost $0.0001 and $0.0011
Two sections in one answerBoth passedClaude Haiku 5.5, 2.5×took 0.8 s and 2.1 sClaude Haiku 5.5, 12.8×cost $0.0001 and $0.0011
A later amendment changes the answerBoth passedClaude Haiku 5.5, 1.1×took 1.4 s and 1.5 sClaude Haiku 5.5, 10.6×cost $0.0001 and $0.0014
A question the handbook doesn't answerBoth passedClaude Haiku 5.5, 1.3×took 0.8 s and 1.0 sClaude Haiku 5.5, 13.5×cost $0.0001 and $0.0011
Pick the tool and work out the dateBoth passedClaude Haiku 4.5, 1.2×took 1.0 s and 0.8 sClaude Haiku 5.5, 14.8×cost $0.0001 and $0.0010
Convert a currencyBoth passedClosetook 0.8 s and 0.7 sClaude Haiku 5.5, 15.0×cost $0.0001 and $0.0009
Book a meeting from a sentenceBoth passedClaude Haiku 4.5, 2.0×took 2.1 s and 1.0 sClaude Haiku 5.5, 8.0×cost $0.0002 and $0.0012
Search, but don't bookBoth passedClaude Haiku 5.5, 1.3×took 1.1 s and 1.4 sClaude Haiku 5.5, 15.8×cost $0.0001 and $0.0011
Two calls with a unit conversionBoth passedClaude Haiku 4.5, 1.7×took 1.7 s and 1.0 sClaude Haiku 5.5, 8.7×cost $0.0001 and $0.0012

Claude Haiku 5.5 vs Claude Haiku 4.5 in our test runs

Claude Haiku 5.5 and Claude Haiku 4.5 on the same prompts, job by job: how many replies passed their check.

Based on 100 of our test runs on , through OpenRouter with the app's own prompt and settings.

Claude Haiku 5.5 and Claude Haiku 4.5 in our test runs, job by job
JobClaude Haiku 5.5Claude Haiku 4.5
Coding5 of 53 of 5
Writing3 of 54 of 5
Math5 of 55 of 5
Summarization3 of 54 of 5
Data analysis5 of 54 of 5
Customer support5 of 54 of 5
Translation5 of 55 of 5
SQL5 of 55 of 5
RAG and answering from documents5 of 55 of 5
Agents and tool use5 of 55 of 5
All jobs46 of 50 · 1 message each · $0.0002 a reply44 of 50 · 1 message each · $0.0016 a reply

The same prompts, sent to both through llmwise's own pipeline and checked the same way: five for each job. A bold count is the one that passed more. Messages: what each reply counted as on Pro. Cost: what OpenRouter charged us per reply, on average. Every prompt, and how it's scored.

Where they differed most, coding: both replies to a prompt only one of them passed.

A hard prompt: “Evaluate an arithmetic expression, no eval”, in full.

  • Claude Haiku 5.5

    Passed: All 15 tests passed.

    ```javascript
    function evaluate(expression) {
      const src = String(expression);
    
      // Tokenize: numbers, operators, parentheses. Spaces are skipped.
      const tokens = [];
      for (let i = 0; i < src.length; ) {
        const c = src[i];
        if (/\s/.test(c)) {
          i++;
          continue;
        }
        const m = /^(\d+(?:\.\d+)?|\.\d+)/.exec(src.slice(i));
        if (m) {
          tokens.push({ type: 'num', value: Number(m[1]) });
          i += m[1].length;
          continue;
        }
        if ('+-*/^()'.includes(c)) {
          tokens.push({ type: c });
          i++;
          continue;
        }
        throw new Error(`Unexpected character "${c}" at position ${i}`);
      }
    
      // Recursive descent parser.
      // expr    := term (('+' | '-') term)*
      // term    := unary (('*' | '/') unary)*
      // unary   := '-' unary | power
      // power   := primary ('^' unary)?      (right-associative, exponent may be signed)
      // primary := number | '(' expr ')'
      let p = 0;
      const peek = () => (p < tokens.length ? tokens[p].type : null);
      const expect = (type) => {
        if (peek() !== type) {
          throw new Error(`Expected "${type}" at token ${p + 1}`);
        }
        p++;
      };
    
      function parseExpr() {
        let v = parseTerm();
        while (peek() === '+' || peek() === '-') {
          const op = tokens[p++].type;
          const r = parseTerm();
          v = op === '+' ? v + r : v - r;
        }
        return v;
      }
    
      function parseTerm() {
        let v = parseUnary();
        while (peek() === '*' || peek() === '/') {
          const op = tokens[p++].type;
          const r = parseUnary();
          v = op === '*' ? v * r : v / r;
        }
        return v;
      }
    
      function parseUnary() {
        if (peek() === '-') {
          p++;
          return -parseUnary();
        }
        return parsePower();
      }
    
      function parsePower() {
        const base = parsePrimary();
        if (peek() === '^') {
          p++;
          return base ** parseUnary();
        }
        return base;
      }
    
      function parsePrimary() {
        if (peek() === 'num') {
          return tokens[p++].value;
        }
        if (peek() === '(') {
          p++;
          const v = parseExpr();
          expect(')');
          return v;
        }
        if (peek() === null) {
          throw new Error('Unexpected end of expression');
        }
        throw new Error(`Unexpected token "${tokens[p].type}" at token ${p + 1}`);
      }
    
      if (tokens.length === 0) {
        throw new Error('Empty expression');
      }
      const result = parseExpr();
      if (p !== tokens.length) {
        throw new Error(`Unexpected token "${tokens[p].type}" at token ${p + 1}`);
      }
      return result;
    }
    ```

    1,017 tokens in, 2,412 out (1,349 of them reasoning) · 9.3 s · $0.0013 · 1 message on Pro · answered by anthropic/claude-haiku-5.5 via Anthropic ·

  • Claude Haiku 4.5

    Failed: 14 of 15 tests passed. First failure: evaluate("-2 ^ 2"): Expected values to be strictly deep-equal:

    ```javascript
    function evaluate(expression) {
      const tokens = tokenize(expression);
      const parser = new Parser(tokens);
      const result = parser.parseExpression();
      if (parser.pos !== tokens.length) {
        throw new Error("Unexpected token after expression");
      }
      return result;
    }
    
    function tokenize(expression) {
      const tokens = [];
      let i = 0;
      while (i < expression.length) {
        if (/\s/.test(expression[i])) {
          i++;
          continue;
        }
        if (/\d/.test(expression[i])) {
          let num = "";
          while (i < expression.length && /[\d.]/.test(expression[i])) {
            num += expression[i];
            i++;
          }
          if ((num.match(/\./g) || []).length > 1) {
            throw new Error("Invalid number");
          }
          tokens.push({ type: "number", value: parseFloat(num) });
        } else if (/[+\-*/^()]/.test(expression[i])) {
          tokens.push({ type: expression[i], value: expression[i] });
          i++;
        } else {
          throw new Error(`Invalid character: ${expression[i]}`);
        }
      }
      return tokens;
    }
    
    class Parser {
      constructor(tokens) {
        this.tokens = tokens;
        this.pos = 0;
      }
    
      parseExpression() {
        return this.parseAdditive();
      }
    
      parseAdditive() {
        let left = this.parseMultiplicative();
        while (this.pos < this.tokens.length && (this.tokens[this.pos].type === "+" || this.tokens[this.pos].type === "-")) {
          const op = this.tokens[this.pos].type;
          this.pos++;
          const right = this.parseMultiplicative();
          left = op === "+" ? left + right : left - right;
        }
        return left;
      }
    
      parseMultiplicative() {
        let left = this.parseExponentiation();
        while (this.pos < this.tokens.length && (this.tokens[this.pos].type === "*" || this.tokens[this.pos].type === "/")) {
          const op = this.tokens[this.pos].type;
          this.pos++;
          const right = this.parseExponentiation();
          left = op === "*" ? left * right : left / right;
        }
        return left;
      }
    
      parseExponentiation() {
        let left = this.parseUnary();
        if (this.pos < this.tokens.length && this.tokens[this.pos].type === "^") {
          this.pos++;
          const right = this.parseExponentiation();
          left = Math.pow(left, right);
        }
        return left;
      }
    
      parseUnary() {
        if (this.pos < this.tokens.length && this.tokens[this.pos].type === "-") {
          this.pos++;
          return -this.parseUnary();
        }
        if (this.pos < this.tokens.length && this.tokens[this.pos].type === "+") {
          this.pos++;
          return this.parseUnary();
        }
        return this.parsePrimary();
      }
    
      parsePrimary() {
        if (this.pos >= this.tokens.length) {
          throw new Error("Unexpected end of expression");
        }
        if (this.tokens[this.pos].type === "number") {
          return this.tokens[this.pos++].value;
        }
        if (this.tokens[this.pos].type === "(") {
          this.pos++;
          const result = this.parseExpression();
          if (this.pos >= this.tokens.length || this.tokens[this.pos].type !== ")") {
            throw new Error("Missing closing parenthesis");
          }
          this.pos++;
          return result;
        }
        throw new Error(`Unexpected token: ${this.tokens[this.pos].value}`);
      }
    }
    ```

    788 tokens in, 1,017 out · 5.7 s · $0.0059 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Anthropic ·

Claude Haiku 5.5 and Claude Haiku 4.5 on every plan

Each plan's messages on it. Every plan is month to month: cancel in two clicks.

Messages on Claude Haiku 5.5 and Claude Haiku 4.5, plan by plan
PlanPriceClaude Haiku 5.5Claude Haiku 4.5
Plus$10 a month30 a dayUp to 125 a month
Pro$20 a month60 a dayUp to 250 a month
Max$50 a month120 a dayUp to 800 a month
Ultra$100 a month200 a dayUp to 1,800 a month
Studio$200 a month200 a dayUp to 4,000 a month

Prices don't include tax, which is added where it applies and shown before you pay. A paid plan's month is one allowance shared by every model, so each monthly count is the most you get if all of it goes to that model. It renews each billing period; everyday models refill daily at 00:00 UTC. Long chats count more per reply. How pricing works.

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.

What differs

  • Messages on Pro

    Claude Haiku 5.5 is an everyday model (60 messages a day on Pro); Claude Haiku 4.5 draws on the monthly allowance (up to 250 messages a month on Pro).

  • Context window

    Claude Haiku 5.5 takes up to 1M tokens; Claude Haiku 4.5 up to 200K tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.

  • Images and PDFs

    Both read images. Both take a PDF as the whole file, pages and all.

  • Plans

    Claude Haiku 5.5 and Claude Haiku 4.5 are both on Plus and Pro.

  • Where messages go

    Both are Claude models, served the same way. Sent to Anthropic directly.

Fact by fact

Claude Haiku 5.5 and Claude Haiku 4.5, fact by fact
FactClaude Haiku 5.5Claude Haiku 4.5
Context window1M tokens200K tokens
Reads imagesYesYes
PDFsWhole fileWhole file
ReasoningYesNo
API price (October 2026)$0.10 in / $0.50 out per million tokens$1.00 in / $5.00 out per million tokens
A typical message at API prices (4,000 tokens in, 700 out)$0.0008$0.0075
A $10 top-up addsNothing: an everyday model's count is daily200 messages
Where a message goesSent to Anthropic directly.Sent to Anthropic directly.
If the provider failsIf Anthropic fails before the reply starts (an overload, a server error, a dropped connection), llmwise sends the same request to Claude Haiku 5.5 through OpenRouter instead.If Anthropic fails before the reply starts (an overload, a server error, a dropped connection), llmwise sends the same request to Claude Haiku 4.5 through OpenRouter instead.
Anthropic's safety fallbackDoesn't applyDoesn't apply
API prices are what our model catalog lists (Claude: Anthropic's list price). In llmwise you pay per message, not per token: the counts above are what you get.

Claude Haiku 5.5 or Claude Haiku 4.5?

From the facts above and our test runs: the rest is how their answers suit your work, which one chat can show you.

  • Pick Claude Haiku 5.5: its messages come from the daily count (60 messages a day on Pro), so they leave the monthly allowance for bigger models.

Each model's page, the families, and other pairs

Claude Haiku 5.5 vs Claude Haiku 4.5 is one pair of models. The page below covers the whole family.

Questions

Is Claude Haiku 5.5 or Claude Haiku 4.5 cheaper in llmwise?

Claude Haiku 5.5 is an everyday model (60 messages a day on Pro); Claude Haiku 4.5 draws on the monthly allowance (up to 250 messages a month on Pro). Every paid plan's monthly allowance is shared by the models on it, so each count is the most you get if it all goes to that model.

Can I try Claude Haiku 5.5 and Claude Haiku 4.5 for free?

No: there's no free plan or trial. Claude Haiku 5.5 and Claude Haiku 4.5 are both on Plus and Pro. Paid monthly from the day you start. Cancel in two clicks in Settings → Billing.

Which has the bigger context window, Claude Haiku 5.5 or Claude Haiku 4.5?

Claude Haiku 5.5: 1M tokens, against 200K tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.

Can I use Claude Haiku 5.5 and Claude Haiku 4.5 in the same chat?

Yes. Pick Claude Haiku 5.5 for one message and Claude Haiku 4.5 for the next; the second sees the whole chat, including the first one's answer.

Which did better in your test runs, Claude Haiku 5.5 or Claude Haiku 4.5?

On the same 50 prompts, run on November 18, 2026, Claude Haiku 5.5 passed 46 and Claude Haiku 4.5 passed 44. The table on this page has each job, and every reply is published.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.

See what a message costs before you send it. Every plan is month to month: cancel in two clicks. Sign in with an email link, no password.