Skip to content

Claude Fable 5.2 · Coding

Claude Fable 5.2 for coding: our test runs

Claude Fable 5.2 ran our 5 coding prompts in llmwise, through the same pipeline your messages take. Here's every reply as it came, whether it passed, what it cost and what it counts as on Pro, next to the other models.

Based on 5 of our test runs on , through OpenRouter with the app's own prompt and settings.

Short answer

In our test runs on November 17, 2026, Claude Fable 5.2 passed 5 of 5 of our coding prompts (both hard ones): tied 1st of 27 models. Each reply counted as 1 message on Pro, where it has 31 a month, and cost $0.0249 to run on average.

Prompt by prompt

Each of the 5 coding prompts, and how Claude Fable 5.2's reply did.

Claude Fable 5.2 on each coding prompt
PromptResultCostTime
Turn a title into a URL slugEverydayPassed: All 7 tests passed.$0.01583.3 s
Parse a duration like “1h 30m”EverydayPassed: All 10 tests passed.$0.01454.1 s
Merge overlapping intervalsEverydayPassed: All 7 tests passed.$0.01123.1 s
Evaluate an arithmetic expression, no evalHardPassed: All 15 tests passed.$0.059312.2 s
Parse CSV with quoted fieldsHardPassed: All 8 tests passed.$0.02335.2 s

Every coding reply from Claude Fable 5.2

  1. Turn a title into a URL slug

    An everyday prompt: “Turn a title into a URL slug”, in full.

    Passed: All 7 tests passed.

    ```javascript
    function slugify(title) {
      return String(title)
        .normalize("NFD")                 // split accented letters into base letter + combining mark
        .replace(/[\u0300-\u036f]/g, "")  // remove the combining marks
        .toLowerCase()
        .replace(/[^a-z0-9]+/g, "-")      // collapse runs of other characters into one hyphen
        .replace(/^-+|-+$/g, "");         // trim hyphens from both ends
    }
    ```

    925 tokens in, 167 out · 3.3 s · $0.0158 · 1 message on Pro · answered by anthropic/claude-fable-5.2 via Google ·

  2. Parse a duration like “1h 30m”

    An everyday prompt: “Parse a duration like “1h 30m””, in full.

    Passed: All 10 tests passed.

    ```javascript
    function parseDuration(text) {
      if (typeof text !== "string") return null;
    
      const match = /^[ \t]*(?:(\d+)h)?[ \t]*(?:(\d+)m)?[ \t]*(?:(\d+)s)?[ \t]*$/.exec(text);
      if (!match) return null;
    
      const [, h, m, s] = match;
    
      // Reject inputs with no parts at all (e.g. "" or "   ").
      if (h === undefined && m === undefined && s === undefined) return null;
    
      return Number(h ?? 0) * 3600 + Number(m ?? 0) * 60 + Number(s ?? 0);
    }
    ```
    
    The regex enforces the h, m, s order with each part appearing at most once, so inputs like `"5m 1h"`, `"1x"`, and `"h"` return `null`. Values aren't capped, so `"90m"` gives 5400.

    880 tokens in, 307 out · 4.1 s · $0.0145 · 1 message on Pro · answered by anthropic/claude-fable-5.2 via Google ·

  3. Merge overlapping intervals

    An everyday prompt: “Merge overlapping intervals”, in full.

    Passed: All 7 tests passed.

    ```javascript
    function mergeIntervals(intervals) {
      // Copy each pair so neither the input array nor its inner arrays are mutated
      const sorted = intervals
        .map(([start, end]) => [start, end])
        .sort((a, b) => a[0] - b[0] || a[1] - b[1]);
    
      const result = [];
      for (const [start, end] of sorted) {
        const last = result[result.length - 1];
        if (last && start <= last[1]) {
          // Overlapping or touching: extend the previous interval
          if (end > last[1]) last[1] = end;
        } else {
          result.push([start, end]);
        }
      }
      return result;
    }
    ```

    875 tokens in, 226 out · 3.1 s · $0.0112 · 1 message on Pro · answered by anthropic/claude-fable-5.2 via Google ·

  4. Evaluate an arithmetic expression, no eval

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

    Passed: All 15 tests passed.

    ```javascript
    function evaluate(expression) {
      if (typeof expression !== 'string') {
        throw new Error('Expression must be a string');
      }
    
      const s = expression;
      let pos = 0;
    
      function skipSpaces() {
        while (pos < s.length && /\s/.test(s[pos])) pos++;
      }
    
      function peek() {
        skipSpaces();
        return pos < s.length ? s[pos] : null;
      }
    
      function isDigit(ch) {
        return ch !== undefined && ch >= '0' && ch <= '9';
      }
    
      // expr := term (('+' | '-') term)*
      function parseExpr() {
        let value = parseTerm();
        for (;;) {
          const ch = peek();
          if (ch === '+') {
            pos++;
            value += parseTerm();
          } else if (ch === '-') {
            pos++;
            value -= parseTerm();
          } else {
            return value;
          }
        }
      }
    
      // term := unary (('*' | '/') unary)*
      function parseTerm() {
        let value = parseUnary();
        for (;;) {
          const ch = peek();
          if (ch === '*') {
            pos++;
            value *= parseUnary();
          } else if (ch === '/') {
            pos++;
            value /= parseUnary();
          } else {
            return value;
          }
        }
      }
    
      // unary := '-' unary | power
      function parseUnary() {
        if (peek() === '-') {
          pos++;
          return -parseUnary();
        }
        return parsePower();
      }
    
      // power := primary ('^' unary)?   (right-associative, exponent may be negated)
      function parsePower() {
        const base = parsePrimary();
        if (peek() === '^') {
          pos++;
          const exponent = parseUnary();
          return Math.pow(base, exponent);
        }
        return base;
      }
    
      // primary := number | '(' expr ')'
      function parsePrimary() {
        const ch = peek();
        if (ch === null) {
          throw new Error('Unexpected end of expression');
        }
        if (ch === '(') {
          pos++;
          const value = parseExpr();
          if (peek() !== ')') {
            throw new Error('Expected ")" at position ' + pos);
          }
          pos++;
          return value;
        }
        if (isDigit(ch)) {
          const start = pos;
          while (isDigit(s[pos])) pos++;
          if (s[pos] === '.') {
            pos++;
            if (!isDigit(s[pos])) {
              throw new Error('Invalid number at position ' + start);
            }
            while (isDigit(s[pos])) pos++;
          }
          return Number(s.slice(start, pos));
        }
        throw new Error('Unexpected character "' + ch + '" at position ' + pos);
      }
    
      const result = parseExpr();
      if (peek() !== null) {
        throw new Error('Unexpected character "' + s[pos] + '" at position ' + pos);
      }
      return result;
    }
    ```

    1,016 tokens in, 1,393 out (374 of them reasoning) · 12.2 s · $0.0593 · 1 message on Pro · answered by anthropic/claude-fable-5.2 via Google ·

  5. Parse CSV with quoted fields

    A hard prompt: “Parse CSV with quoted fields”, in full.

    Passed: All 8 tests passed.

    ```javascript
    function parseCsv(text) {
      const rows = [];
      let row = [];
      let field = '';
      let inQuotes = false;
      let fieldStart = true;   // at the very start of a field
      let pending = false;     // true if the current row has any content (fields or a pending field)
      const n = text.length;
      let i = 0;
    
      while (i < n) {
        const ch = text[i];
    
        if (inQuotes) {
          if (ch === '"') {
            if (text[i + 1] === '"') {
              field += '"';
              i += 2;
            } else {
              inQuotes = false;
              i++;
            }
          } else {
            field += ch;
            i++;
          }
          continue;
        }
    
        if (ch === '"' && fieldStart) {
          inQuotes = true;
          fieldStart = false;
          pending = true;
          i++;
        } else if (ch === ',') {
          row.push(field);
          field = '';
          fieldStart = true;
          pending = true;
          i++;
        } else if (ch === '\n' || (ch === '\r' && text[i + 1] === '\n')) {
          row.push(field);
          rows.push(row);
          row = [];
          field = '';
          fieldStart = true;
          pending = false;
          i += ch === '\r' ? 2 : 1;
        } else {
          field += ch;
          fieldStart = false;
          pending = true;
          i++;
        }
      }
    
      if (pending) {
        row.push(field);
        rows.push(row);
      }
    
      return rows;
    }
    ```

    923 tokens in, 516 out · 5.2 s · $0.0233 · 1 message on Pro · answered by anthropic/claude-fable-5.2 via Google ·

How it compares on coding

Claude Fable 5.2 beside the job's picks among every model. 25 models share the hard-problems pick, Claude Fable 5.2 among them.

Claude Fable 5.2 beside other models on coding
ModelPassedOn ProCost per reply
Gemini 3.8 FlashOur pick: hard problems (shared), best value5 of 5250 a month on Pro$0.0011
GPT-6 LunaOur pick: hard problems (shared), everyday5 of 560 a day on Pro$0.0001
Claude Fable 5.2Our pick: hard problems (shared)5 of 531 a month on Pro$0.0249
On Pro: Pro's count on each model. Cost: what OpenRouter charged us per reply, on average.

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.

Claude Fable 5.2 next to each model on coding

Claude Fable 5.2 against each model on the same 5 coding prompts. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.

  • Against Claude Fable 5.1: Claude Fable 5.2 answered within 10% of its time (4.1 s to 4.5 s), cost 1.3× less, and passed 5 of 5 to its 5.

  • Against Claude Opus 5.5: Claude Fable 5.2 answered 1.4× sooner (4.1 s to 5.7 s), cost 1.4× more, and passed 5 of 5 to its 5.

  • Against Claude Sonnet 5.5: Claude Fable 5.2 answered 1.7× later (4.1 s to 2.5 s), cost 4.9× more, and passed 5 of 5 to its 5.

  • Against Claude Sonnet 5: Claude Fable 5.2 answered 1.6× later (4.1 s to 2.6 s), cost 2.5× more, and passed 5 of 5 to its 5.

  • Against Claude Haiku 5.5: Claude Fable 5.2 answered 3.2× later (4.1 s to 1.3 s), cost 62.7× more, and passed 5 of 5 to its 5.

  • Against Claude Haiku 4.5: Claude Fable 5.2 answered 1.9× later (4.1 s to 2.2 s), cost 8.3× more, and passed 5 of 5 to its 3. Only Claude Fable 5.2 passed Evaluate an arithmetic expression, no eval; Parse CSV with quoted fields.

  • Against GPT-6.2 Astra: Claude Fable 5.2 answered 1.2× later (4.1 s to 3.5 s), cost 1.9× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Astra: Claude Fable 5.2 answered 1.4× sooner (4.1 s to 5.8 s), cost 2.3× more, and passed 5 of 5 to its 5.

  • Against GPT-6.2 Sol: Claude Fable 5.2 answered within 10% of its time (4.1 s to 3.8 s), cost 6.7× more, and passed 5 of 5 to its 5.

  • Against GPT-6.1 Sol: Claude Fable 5.2 answered 1.9× sooner (4.1 s to 7.9 s), cost 6.1× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Sol: Claude Fable 5.2 answered 1.2× sooner (4.1 s to 4.8 s), cost 8.8× more, and passed 5 of 5 to its 5.

  • Against GPT-6.1 Terra: Claude Fable 5.2 answered 1.6× sooner (4.1 s to 6.8 s), cost 18.7× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Luna: Claude Fable 5.2 answered 1.8× sooner (4.1 s to 7.3 s), cost 176.0× more, and passed 5 of 5 to its 5.

  • Against Gemini 4 Carbon: Claude Fable 5.2 answered 8.8× sooner (4.1 s to 36.4 s), cost 1.8× more, and passed 5 of 5 to its 5.

  • Against Gemini 4 Argon: Claude Fable 5.2 answered 1.8× sooner (4.1 s to 7.6 s), cost 9.6× more, and passed 5 of 5 to its 5.

  • Against Gemini 4 Hydrogen: Claude Fable 5.2 answered 1.7× sooner (4.1 s to 6.9 s), cost 4.1× more, and passed 5 of 5 to its 5.

  • Against Gemini 3.1 Pro (preview): Claude Fable 5.2 answered 3.8× sooner (4.1 s to 15.7 s), cost 1.9× more, and passed 5 of 5 to its 5.

  • Against Gemini 3.8 Flash: Claude Fable 5.2 answered 1.3× later (4.1 s to 3.3 s), cost 22.4× more, and passed 5 of 5 to its 5.

  • Against DeepSeek V4.1 Flash: Claude Fable 5.2 answered 1.2× later (4.1 s to 3.5 s), cost 17.0× more, and passed 5 of 5 to its 5.

  • Against DeepSeek V4 Pro: Claude Fable 5.2 answered 2.7× sooner (4.1 s to 11.3 s), cost 2.1× more, and passed 5 of 5 to its 5.

  • Against Grok 4.7: Claude Fable 5.2 answered 7.6× sooner (4.1 s to 31.2 s), cost within 10% of it, and passed 5 of 5 to its 5.

  • Against Kimi K3.1 Pro: Claude Fable 5.2 answered 1.9× sooner (4.1 s to 7.8 s), cost 3.3× more, and passed 5 of 5 to its 5.

  • Against Kimi K3: Claude Fable 5.2 answered 1.5× sooner (4.1 s to 6.2 s), cost 3.4× more, and passed 5 of 5 to its 5.

  • Against GLM 5.3: Claude Fable 5.2 answered 2.8× later (4.1 s to 1.5 s), cost 10.9× more, and passed 5 of 5 to its 3. Only Claude Fable 5.2 passed Parse a duration like “1h 30m”; Parse CSV with quoted fields.

  • Against GLM 5.3 Flash: Claude Fable 5.2 answered 3.9× later (4.1 s to 1.1 s), cost 95.3× more, and passed 5 of 5 to its 5.

  • Against Mistral Large 4: Claude Fable 5.2 answered 11.4× sooner (4.1 s to 46.9 s), cost 3.4× more, and passed 5 of 5 to its 5.

How these runs were done

Tests. Automatic. The function runs against the prompt's tests in a separate Node.js process with a time limit and no file, network or child-process access; it passes when every test passes.

How the runs were done, and every coding prompt.

Claude Fable 5.2, and coding, elsewhere

Questions

Is Claude Fable 5.2 good for coding?

In our test runs it passed 5 of 5 coding prompts, tied 1st of the 27 models in llmwise. Every reply is on this page, so you can judge them yourself.

How many of my messages does a coding reply from Claude Fable 5.2 use?

1 message each on Pro, where it has 31 a month on Pro. The price of a message is fixed and shown before you send it, however long the reply.

How were these runs done?

The same way for every model: each prompt sent through llmwise's own pipeline, each reply checked the same way. The methods page has every prompt and how each is scored.

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.