Model vs model
Claude Haiku 4.5 vs Gemini 3.1 Pro (preview)
Claude Haiku 4.5 and Gemini 3.1 Pro (preview) 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 4.5 page, OpenRouter's Gemini 3.1 Pro page. Updated .
Short answer
Claude Haiku 4.5 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Gemini 3.1 Pro (preview). Beyond that, they read the same files and neither is easier to try. In our test runs, Claude Haiku 4.5 passed 43 of the 50 prompts both answered and Gemini 3.1 Pro (preview) 49; 8 prompts split them, most on coding (3 to 5).
Claude Haiku 4.5 vs Gemini 3.1 Pro (preview), prompt by prompt
Every prompt Claude Haiku 4.5 and Gemini 3.1 Pro (preview) 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 42, only Claude Haiku 4.5 passed 1, only Gemini 3.1 Pro (preview) passed 7, and neither passed 0. Claude Haiku 4.5 answered sooner on 50 of the 50 and Gemini 3.1 Pro (preview) on 0; the rest were within 10% of each other. The 50 replies cost $0.0741 on Claude Haiku 4.5 and $0.5334 on Gemini 3.1 Pro (preview): 7.2× less on Claude Haiku 4.5.
The 8 prompts only one of Claude Haiku 4.5 and Gemini 3.1 Pro (preview) passed
Evaluate an arithmetic expression, no eval (coding): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: 14 of 15 tests passed. First failure: evaluate("-2 ^ 2"): Expected values to be strictly deep-equal: Gemini 3.1 Pro (preview): All 15 tests passed.
Parse CSV with quoted fields (coding): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: 5 of 8 tests passed. First failure: parseCsv('"line1\nline2",end\r\nnext,row\r\n'): Expected values to be strictly deep-equal: Gemini 3.1 Pro (preview): All 8 tests passed.
A product announcement with five rules (writing): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: Graded 3.7 of 5 on average (lowest 3). Gemini 3.1 Pro (preview): Graded 4.0 of 5 on average (lowest 3).
Argue both sides of free buses (writing): Claude Haiku 4.5 passed and Gemini 3.1 Pro (preview) didn't. Claude Haiku 4.5: Graded 4.0 of 5 on average (lowest 4). Gemini 3.1 Pro (preview): Graded 3.7 of 5 on average (lowest 3).
Average order value in August (data analysis): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: Final answer 275.00; expected 300. Gemini 3.1 Pro (preview): Final answer 300.00: right.
Correlation between ad spend and sign-ups (data analysis): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: Final answer 0.99; expected 0.97. Gemini 3.1 Pro (preview): Final answer 0.97: right.
A frustrated customer (customer support): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: Graded 3.3 of 5 on average (lowest 3). Gemini 3.1 Pro (preview): Graded 4.0 of 5 on average (lowest 3).
A message with a planted instruction (customer support): Gemini 3.1 Pro (preview) passed and Claude Haiku 4.5 didn't. Claude Haiku 4.5: Graded 4.3 of 5 on average (lowest 3); but 165 words, over the 150 allowed. Gemini 3.1 Pro (preview): Graded 4.3 of 5 on average (lowest 4).
Job by job, the widest gaps first
Customer support: Claude Haiku 4.5 passed 3 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 3.2× sooner at the median, 2.7 s against 8.7 s. Claude Haiku 4.5 cost 7.2× less, $0.0073 against $0.0529 for the 5 replies. Claude Haiku 4.5's replies ran 32% longer, in tokens of reply, thinking not counted.
Coding: Claude Haiku 4.5 passed 3 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 4.4× sooner at the median, 2.3 s against 10.3 s. Claude Haiku 4.5 cost 7.1× less, $0.0137 against $0.0973 for the 5 replies. Claude Haiku 4.5's replies ran 43% longer, in tokens of reply, thinking not counted.
Data analysis: Claude Haiku 4.5 passed 3 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 3.2× sooner at the median, 2.7 s against 8.6 s. Claude Haiku 4.5 cost 6.2× less, $0.0125 against $0.0773 for the 5 replies. Claude Haiku 4.5's replies ran 14% longer, in tokens of reply, thinking not counted.
Translation: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 5.6× sooner at the median, 1.8 s against 10.3 s. Claude Haiku 4.5 cost 9.4× less, $0.0074 against $0.0699 for the 5 replies. Claude Haiku 4.5's replies ran 99% longer, in tokens of reply, thinking not counted.
Summarization: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 4.8× sooner at the median, 1.7 s against 8.0 s. Claude Haiku 4.5 cost 8.7× less, $0.0051 against $0.0441 for the 5 replies. Claude Haiku 4.5's replies ran 20% longer, in tokens of reply, thinking not counted.
Writing: Claude Haiku 4.5 passed 4 of 5 and Gemini 3.1 Pro (preview) 4 of 5. Claude Haiku 4.5 answered 4.1× sooner at the median, 2.2 s against 9.1 s. Claude Haiku 4.5 cost 8.1× less, $0.0057 against $0.0459 for the 5 replies. Their replies ran to about the same length.
SQL: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 6.0× sooner at the median, 1.2 s against 7.4 s. Claude Haiku 4.5 cost 7.8× less, $0.0049 against $0.0386 for the 5 replies. Claude Haiku 4.5's replies ran 13% longer, in tokens of reply, thinking not counted.
Math: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 3.7× sooner at the median, 2.2 s against 8.3 s. Claude Haiku 4.5 cost 6.5× less, $0.0077 against $0.0501 for the 5 replies. Their replies ran to about the same length.
Agents and tool use: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 4.7× sooner at the median, 1.1 s against 5.3 s. Claude Haiku 4.5 cost 6.0× less, $0.0047 against $0.0284 for the 5 replies. Their replies ran to about the same length.
RAG and answering from documents: Claude Haiku 4.5 passed 5 of 5 and Gemini 3.1 Pro (preview) 5 of 5. Claude Haiku 4.5 answered 4.7× sooner at the median, 1.2 s against 5.6 s. Claude Haiku 4.5 cost 5.7× less, $0.0050 against $0.0290 for the 5 replies. Claude Haiku 4.5's replies ran 176% longer, in tokens of reply, thinking not counted.
All 50 prompts: who passed, who answered sooner, who cost less
| Prompt | Result | Sooner | Cheaper |
|---|---|---|---|
| Turn a title into a URL slug | Both passed | Claude Haiku 4.5, 3.9×took 1.5 s and 6.0 s | Claude Haiku 4.5, 5.8×cost $0.0010 and $0.0056 |
| Parse a duration like “1h 30m” | Both passed | Claude Haiku 4.5, 4.4×took 2.3 s and 10.3 s | Claude Haiku 4.5, 7.9×cost $0.0022 and $0.0172 |
| Merge overlapping intervals | Both passed | Claude Haiku 4.5, 4.4×took 2.0 s and 8.6 s | Claude Haiku 4.5, 6.2×cost $0.0017 and $0.0103 |
| Evaluate an arithmetic expression, no eval | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 2.7×took 7.6 s and 20.6 s | Claude Haiku 4.5, 6.2×cost $0.0056 and $0.0351 |
| Parse CSV with quoted fields | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 5.9×took 3.5 s and 20.6 s | Claude Haiku 4.5, 8.8×cost $0.0033 and $0.0291 |
| Announce a second bakery shop on LinkedIn | Both passed | Claude Haiku 4.5, 3.7×took 2.6 s and 9.5 s | Claude Haiku 4.5, 9.0×cost $0.0013 and $0.0113 |
| Rewrite corporate jargon in plain words | Both passed | Claude Haiku 4.5, 5.5×took 1.6 s and 8.8 s | Claude Haiku 4.5, 12.2×cost $0.0008 and $0.0101 |
| Decline a meeting and offer two times | Both passed | Claude Haiku 4.5, 3.1×took 2.0 s and 6.2 s | Claude Haiku 4.5, 5.9×cost $0.0010 and $0.0058 |
| A product announcement with five rules | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 4.1×took 2.2 s and 9.2 s | Claude Haiku 4.5, 7.8×cost $0.0012 and $0.0095 |
| Argue both sides of free buses | Only Claude Haiku 4.5 | Claude Haiku 4.5, 2.6×took 3.4 s and 9.1 s | Claude Haiku 4.5, 6.6×cost $0.0014 and $0.0092 |
| A discount, then sales tax | Both passed | Claude Haiku 4.5, 3.1×took 1.1 s and 3.6 s | Claude Haiku 4.5, 5.6×cost $0.0007 and $0.0042 |
| Pens at 3 for $4 | Both passed | Claude Haiku 4.5, 4.5×took 2.2 s and 9.6 s | Claude Haiku 4.5, 8.4×cost $0.0014 and $0.0121 |
| Compound interest over three years | Both passed | Claude Haiku 4.5, 1.8×took 3.5 s and 6.2 s | Claude Haiku 4.5, 7.6×cost $0.0013 and $0.0099 |
| Four-digit numbers whose digits sum to 9 | Both passed | Claude Haiku 4.5, 3.4×took 2.9 s and 9.6 s | Claude Haiku 4.5, 6.2×cost $0.0024 and $0.0150 |
| The highest of three dice is a 5 | Both passed | Claude Haiku 4.5, 3.7×took 2.2 s and 8.3 s | Claude Haiku 4.5, 5.0×cost $0.0018 and $0.0090 |
| An article in three bullets | Both passed | Claude Haiku 4.5, 4.4×took 1.8 s and 8.0 s | Claude Haiku 4.5, 8.2×cost $0.0011 and $0.0088 |
| An email thread in one sentence | Both passed | Claude Haiku 4.5, 4.8×took 1.3 s and 6.3 s | Claude Haiku 4.5, 8.0×cost $0.0007 and $0.0058 |
| Decisions and action items from a meeting | Both passed | Claude Haiku 4.5, 4.5×took 1.7 s and 7.6 s | Claude Haiku 4.5, 7.6×cost $0.0011 and $0.0081 |
| A quarterly memo for the CEO | Both passed | Claude Haiku 4.5, 7.3×took 1.5 s and 11.2 s | Claude Haiku 4.5, 11.5×cost $0.0012 and $0.0139 |
| A study with a negative result | Both passed | Claude Haiku 4.5, 2.4×took 3.4 s and 8.2 s | Claude Haiku 4.5, 7.4×cost $0.0010 and $0.0075 |
| The region with the most revenue | Both passed | Claude Haiku 4.5, 3.1×took 2.7 s and 8.5 s | Claude Haiku 4.5, 4.5×cost $0.0026 and $0.0117 |
| Average order value in August | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 4.1×took 2.1 s and 8.6 s | Claude Haiku 4.5, 5.6×cost $0.0023 and $0.0130 |
| Revenue change from July to August | Both passed | Claude Haiku 4.5, 2.1×took 4.2 s and 9.0 s | Claude Haiku 4.5, 6.0×cost $0.0022 and $0.0135 |
| A median, filtered two ways | Both passed | Claude Haiku 4.5, 3.6×took 2.0 s and 7.1 s | Claude Haiku 4.5, 5.0×cost $0.0016 and $0.0082 |
| Correlation between ad spend and sign-ups | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 3.3×took 5.0 s and 16.5 s | Claude Haiku 4.5, 8.4×cost $0.0037 and $0.0310 |
| A late order | Both passed | Claude Haiku 4.5, 4.3×took 2.0 s and 8.7 s | Claude Haiku 4.5, 9.2×cost $0.0014 and $0.0125 |
| A return inside the window | Both passed | Claude Haiku 4.5, 2.9×took 2.4 s and 6.9 s | Claude Haiku 4.5, 5.7×cost $0.0013 and $0.0076 |
| A frustrated customer | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 3.3×took 3.0 s and 9.8 s | Claude Haiku 4.5, 7.0×cost $0.0016 and $0.0110 |
| A refund request outside the window | Both passed | Claude Haiku 4.5, 2.7×took 2.7 s and 7.4 s | Claude Haiku 4.5, 7.9×cost $0.0013 and $0.0103 |
| A message with a planted instruction | Only Gemini 3.1 Pro (preview) | Claude Haiku 4.5, 3.4×took 3.2 s and 10.6 s | Claude Haiku 4.5, 6.5×cost $0.0018 and $0.0116 |
| A delivery message into Spanish | Both passed | Claude Haiku 4.5, 7.8×took 1.5 s and 11.8 s | Claude Haiku 4.5, 13.8×cost $0.0011 and $0.0157 |
| A product description into French | Both passed | Claude Haiku 4.5, 6.5×took 1.4 s and 9.1 s | Claude Haiku 4.5, 12.4×cost $0.0011 and $0.0141 |
| A meeting note into German | Both passed | Claude Haiku 4.5, 7.0×took 1.8 s and 13.0 s | Claude Haiku 4.5, 15.3×cost $0.0010 and $0.0158 |
| Idioms into natural Japanese | Both passed | Claude Haiku 4.5, 3.1×took 3.3 s and 10.3 s | Claude Haiku 4.5, 6.2×cost $0.0019 and $0.0118 |
| A lease clause into Brazilian Portuguese | Both passed | Claude Haiku 4.5, 2.3×took 3.5 s and 8.1 s | Claude Haiku 4.5, 5.8×cost $0.0022 and $0.0125 |
| Customers in one country | Both passed | Claude Haiku 4.5, 4.3×took 1.1 s and 4.9 s | Claude Haiku 4.5, 4.5×cost $0.0007 and $0.0029 |
| Count orders by status | Both passed | Claude Haiku 4.5, 6.1×took 0.8 s and 5.0 s | Claude Haiku 4.5, 6.2×cost $0.0006 and $0.0039 |
| Revenue by category | Both passed | Claude Haiku 4.5, 6.0×took 1.2 s and 7.4 s | Claude Haiku 4.5, 7.0×cost $0.0011 and $0.0074 |
| Every customer, even those without orders | Both passed | Claude Haiku 4.5, 3.7×took 2.1 s and 7.9 s | Claude Haiku 4.5, 8.7×cost $0.0010 and $0.0089 |
| Monthly revenue with a running total | Both passed | Claude Haiku 4.5, 6.8×took 1.6 s and 11.1 s | Claude Haiku 4.5, 9.8×cost $0.0016 and $0.0155 |
| A fact from one section | Both passed | Claude Haiku 4.5, 5.0×took 1.1 s and 5.3 s | Claude Haiku 4.5, 5.0×cost $0.0009 and $0.0045 |
| Core hours and start times | Both passed | Claude Haiku 4.5, 5.1×took 1.1 s and 5.8 s | Claude Haiku 4.5, 5.7×cost $0.0009 and $0.0052 |
| Two sections in one answer | Both passed | Claude Haiku 4.5, 4.2×took 1.3 s and 5.6 s | Claude Haiku 4.5, 4.4×cost $0.0011 and $0.0047 |
| A later amendment changes the answer | Both passed | Claude Haiku 4.5, 4.5×took 1.9 s and 8.6 s | Claude Haiku 4.5, 8.0×cost $0.0012 and $0.0099 |
| A question the handbook doesn't answer | Both passed | Claude Haiku 4.5, 4.7×took 1.2 s and 5.6 s | Claude Haiku 4.5, 5.0×cost $0.0009 and $0.0046 |
| Pick the tool and work out the date | Both passed | Claude Haiku 4.5, 4.0×took 1.1 s and 4.3 s | Claude Haiku 4.5, 5.8×cost $0.0008 and $0.0048 |
| Convert a currency | Both passed | Claude Haiku 4.5, 5.3×took 1.0 s and 5.3 s | Claude Haiku 4.5, 5.3×cost $0.0008 and $0.0044 |
| Book a meeting from a sentence | Both passed | Claude Haiku 4.5, 5.9×took 1.2 s and 7.1 s | Claude Haiku 4.5, 6.5×cost $0.0010 and $0.0068 |
| Search, but don't book | Both passed | Claude Haiku 4.5, 4.2×took 1.1 s and 4.7 s | Claude Haiku 4.5, 6.4×cost $0.0010 and $0.0061 |
| Two calls with a unit conversion | Both passed | Claude Haiku 4.5, 5.6×took 1.2 s and 6.9 s | Claude Haiku 4.5, 6.0×cost $0.0011 and $0.0063 |
Claude Haiku 4.5 vs Gemini 3.1 Pro (preview) in our test runs
Claude Haiku 4.5 and Gemini 3.1 Pro (preview) 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. Updated .
| Job | Claude Haiku 4.5 | Gemini 3.1 Pro (preview) |
|---|---|---|
| Coding | 3 of 5 | 5 of 5 |
| Writing | 4 of 5 | 4 of 5 |
| Math | 5 of 5 | 5 of 5 |
| Summarization | 5 of 5 | 5 of 5 |
| Data analysis | 3 of 5 | 5 of 5 |
| Customer support | 3 of 5 | 5 of 5 |
| Translation | 5 of 5 | 5 of 5 |
| SQL | 5 of 5 | 5 of 5 |
| RAG and answering from documents | 5 of 5 | 5 of 5 |
| Agents and tool use | 5 of 5 | 5 of 5 |
| All jobs | 43 of 50 · 1 message each · $0.0015 a reply | 49 of 50 · 1 message each · $0.0107 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 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 tokens after expression"); } return result; } function tokenize(expression) { const tokens = []; let i = 0; while (i < expression.length) { const ch = expression[i]; if (/\s/.test(ch)) { i++; continue; } if (/\d/.test(ch)) { let num = ''; while (i < expression.length && /[\d.]/.test(expression[i])) { num += expression[i]; i++; } const parsed = parseFloat(num); if (isNaN(parsed)) { throw new Error(`Invalid number: ${num}`); } tokens.push({ type: 'NUMBER', value: parsed }); continue; } if ('+-*/^()'.includes(ch)) { tokens.push({ type: ch, value: ch }); i++; continue; } throw new Error(`Unexpected character: ${ch}`); } return tokens; } class Parser { constructor(tokens) { this.tokens = tokens; this.pos = 0; } peek() { return this.tokens[this.pos]; } consume() { return this.tokens[this.pos++]; } parseExpression() { return this.parseAddSub(); } parseAddSub() { let left = this.parseMulDiv(); while (this.peek() && (this.peek().type === '+' || this.peek().type === '-')) { const op = this.consume().type; const right = this.parseMulDiv(); left = op === '+' ? left + right : left - right; } return left; } parseMulDiv() { let left = this.parsePower(); while (this.peek() && (this.peek().type === '*' || this.peek().type === '/')) { const op = this.consume().type; const right = this.parsePower(); left = op === '*' ? left * right : left / right; } return left; } parsePower() { let left = this.parseUnary(); if (this.peek() && this.peek().type === '^') { this.consume(); const right = this.parsePower(); left = Math.pow(left, right); } return left; } parseUnary() { if (this.peek() && this.peek().type === '-') { this.consume(); return -this.parseUnary(); } if (this.peek() && this.peek().type === '+') { this.consume(); return this.parseUnary(); } return this.parsePrimary(); } parsePrimary() { const token = this.peek(); if (!token) { throw new Error("Unexpected end of expression"); } if (token.type === 'NUMBER') { this.consume(); return token.value; } if (token.type === '(') { this.consume(); const result = this.parseExpression(); if (!this.peek() || this.peek().type !== ')') { throw new Error("Missing closing parenthesis"); } this.consume(); return result; } throw new Error(`Unexpected token: ${token.value}`); } } ```599 tokens in, 1,007 out · 7.6 s · $0.0056 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·
Gemini 3.1 Pro
Passed: All 15 tests passed.
```javascript function evaluate(expression) { let pos = 0; function peek() { while (pos < expression.length && expression[pos] === ' ') { pos++; } return pos < expression.length ? expression[pos] : null; } function parseExpr() { let left = parseTerm(); while (true) { let char = peek(); if (char === '+' || char === '-') { pos++; let right = parseTerm(); if (char === '+') left += right; else left -= right; } else { break; } } return left; } function parseTerm() { let left = parseFactor(); while (true) { let char = peek(); if (char === '*' || char === '/') { pos++; let right = parseFactor(); if (char === '*') left *= right; else left /= right; } else { break; } } return left; } function parseFactor() { let char = peek(); if (char === '-') { pos++; return -parseFactor(); } return parsePower(); } function parsePower() { let left = parsePrimary(); let char = peek(); if (char === '^') { pos++; let right = parseFactor(); return Math.pow(left, right); } return left; } function parsePrimary() { let char = peek(); if (char === null) { throw new Error("Unexpected end of expression"); } if (char === '(') { pos++; let val = parseExpr(); if (peek() !== ')') { throw new Error("Missing closing parenthesis"); } pos++; return val; } let match = expression.slice(pos).match(/^\d+(\.\d+)?/); if (!match) { throw new Error(`Invalid character at position ${pos}`); } pos += match[0].length; return parseFloat(match[0]); } let result = parseExpr(); if (peek() !== null) { throw new Error(`Unexpected character at position ${pos}`); } return result; } ```587 tokens in, 2,831 out (2,262 of them reasoning) · 20.6 s · $0.0351 · 1 message on Pro · answered by google/gemini-3.1-pro-preview via Google ·
Claude Haiku 4.5 and Gemini 3.1 Pro (preview) on every plan
Whether the one-time free trial reaches each model, then each paid plan's messages on it.
| Plan | Price | Claude Haiku 4.5 | Gemini 3.1 Pro (preview) |
|---|---|---|---|
| Free | $0 | In the one-time trial of 5 messages | In the one-time trial of 5 messages |
| Pro | $20 a month | Up to 250 a month | Up to 125 a month |
| Max | $50 a month | Up to 800 a month | Up to 400 a month |
| Ultra | $100 a month | Up to 1,800 a month | Up to 900 a month |
| Studio | $200 a month | Up to 4,000 a month | Up to 2,000 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: 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 4.5 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Gemini 3.1 Pro (preview).
Context window
Claude Haiku 4.5 takes up to 200K tokens; Gemini 3.1 Pro (preview) up to 1.05M 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.
On Free
Both are in the free trial.
Where messages go
Claude Haiku 4.5: Sent to Anthropic directly. Gemini 3.1 Pro (preview): Sent to Google directly.
Fact by fact
| Fact | Claude Haiku 4.5 | Gemini 3.1 Pro (preview) |
|---|---|---|
| Context window | 200K tokens | 1.05M tokens |
| Reads images | Yes | Yes |
| PDFs | Whole file | Whole file |
| Reasoning | No | Yes |
| API price (September 2026) | $1.00 in / $5.00 out per million tokens | $2.00 in / $12.00 out per million tokens |
| A typical message at API prices (4,000 tokens in, 700 out) | $0.0075 | $0.0164 |
| A $10 top-up adds | 200 messages | 100 messages |
| Where a message goes | Sent to Anthropic directly. | Sent to Google directly. |
| If the provider fails | 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. | If Google fails before the reply starts (an overload, a server error, a dropped connection), llmwise sends the same request to Gemini 3.1 Pro (preview) through OpenRouter instead. |
| Anthropic's safety fallback | Doesn't apply | Doesn't apply |
Claude Haiku 4.5 or Gemini 3.1 Pro (preview)?
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 4.5: more messages for the money, up to 250 messages a month on Pro.
Each model's page, the families, and other pairs
Claude Haiku 4.5 vs Gemini 3.1 Pro (preview) is one pair of models. The page below covers the whole families.
- Claude Haiku 4.5: price, limits and messages on every plan
- Gemini 3.1 Pro: price, limits and messages on every plan
- Claude vs Gemini
- Claude Fable 5.1 vs Claude Haiku 4.5
- Claude Opus 5.5 vs Claude Haiku 4.5
- Claude Sonnet 5 vs Claude Haiku 4.5
- Claude Opus 5.5 vs Gemini 3.1 Pro (preview)
- Claude Sonnet 5 vs Gemini 3.1 Pro (preview)
- GPT-6 Sol vs Gemini 3.1 Pro (preview)
- Every model-vs-model page
Claude Haiku 4.5 is a Claude model; Gemini 3.1 Pro (preview) is a Gemini model.
Questions
Is Claude Haiku 4.5 or Gemini 3.1 Pro (preview) cheaper in llmwise?
Claude Haiku 4.5 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Gemini 3.1 Pro (preview). Every paid plan's monthly allowance is shared by all models, so each count is the most you get if it all goes to that model.
Can I try Claude Haiku 4.5 and Gemini 3.1 Pro (preview) for free?
Yes: both are in the free trial of 5 messages.
Which has the bigger context window, Claude Haiku 4.5 or Gemini 3.1 Pro (preview)?
Gemini 3.1 Pro (preview): 1.05M 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 4.5 and Gemini 3.1 Pro (preview) in the same chat?
Yes. Pick Claude Haiku 4.5 for one message and Gemini 3.1 Pro (preview) for the next; the second sees the whole chat, including the first one's answer.
Which did better in your test runs, Claude Haiku 4.5 or Gemini 3.1 Pro (preview)?
On the same 50 prompts, run on September 27, 2026, Claude Haiku 4.5 passed 43 and Gemini 3.1 Pro (preview) passed 49. The table on this page has each job, and every reply is published.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.