Compare / Claude Haiku 4.5 vs Claude Haiku 5.5
Claude Haiku 4.5 vs Claude Haiku 5.5
Which is cheaper, which is more capable, and which does better at each kind of work, from 92 results on 6 sources that measured both models.
Of the 92 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331).
Claude Haiku 4.5 or Claude Haiku 5.5? The short answer
Claude Haiku 5.5 is the newer version and costs less, and it scores higher (Spring Prompt overall 38 against 7): upgrade.
- Spring Prompt overall: Claude Haiku 4.5 7, Claude Haiku 5.5 38 (out of 100; how the score works).
- Claude Haiku 5.5 is cheaper to run: $0.001 against $0.006 per product listing on CatalogBench (5.1 times less).
- Claude Haiku 5.5 scores higher on the Artificial Analysis Intelligence Index: 43.4 against 16.9.
- Claude Haiku 5.5 is newer: released 7 Oct 2026, against 15 Oct 2025 for Claude Haiku 4.5.
- On our own benchmarks, on the results both models have: Claude Haiku 5.5 does better for product listings, decks from an analysis, user surveys and marketing planning.
- On third-party benchmarks, on the results both models have: Claude Haiku 5.5 does better for agents and tool use and reasoning and knowledge.
Head to head
Every fact side by side
| Claude Haiku 4.5 | Claude Haiku 5.5 | |
|---|---|---|
| Price per million tokens | $1.00 in · $5.00 out | $0.10 in · $0.50 out |
| Blended (3 in : 1 out) | $2.00 | $0.20 |
| Cheapest host, blended | $2.00 (Amazon Bedrock) | $0.20 (Amazon Bedrock) |
| Cost per product listing (CatalogBench) | $0.006 | $0.001 |
| Cached input, per million | $0.10 cached input | – |
| Intelligence Index | 16.9 | 43.4 |
| Spring Prompt overall | 7 | 38 |
| Output speed, tokens a second | – | 240 |
| Context window | 200,000 tokens | 1,000,000 tokens |
| Released | 15 Oct 2025 | 7 Oct 2026 (newer) |
| Weights | Closed | Closed |
| Developer based in | the US | the US |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Their slides, side by side
The same brief for both: Manchester store investment review for Hearthside Coffee, an invented company. First slide of each deck from DeckBench.
From our research
- Claude Haiku 5.5 Is Far Better Than Haiku 4.5 at a Tenth of the Price. GPT-6 Luna Still Beats It at Business Work. 8 Oct 2026
- AI Checkpoint: Sol’s new fast lane 9 Oct 2026
- Muse Ember review: GPT-6 Luna's price, better at analysis, and no Gemini Flash against a clock 10 Nov 2026
- How to Test AI Hallucinations in Business Content 20 Oct 2026
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Claude Haiku 4.5 | Claude Haiku 5.5 |
|---|---|---|
| BulletBench measured by Spring Prompt · 3 results | ||
| Cost per game · Blitz 3+2US dollars, lower is better 6 · 4 of 23 measured | $0.0318 | $0.0164 |
| Cost per game · Bullet 60sUS dollars, lower is better 45 · 23 of 62 measured | $0.0163 | $0.0019no reasoning |
| Ladder Elo · Bullet 60sladder Elo, higher is better 41 · 50 of 62 measured | 445 | 310no reasoning |
| CatalogBench measured by Spring Prompt · 3 results | ||
| Cost per productUS dollars, lower is better 12 · 4 of 36 measured | $0.0061 | $0.0012 |
| Reliably publish-ready% of products, higher is better 32 · 26 of 35 measured | 1.8% | 26.8% |
| Reliably publish-ready · sales brief% of products, higher is better 30 · 20 of 35 measured | 0.0% | 10.7% |
| DeckBench measured by Spring Prompt · 3 results | ||
| Cost per deckUS dollars, lower is better 12 · 4 of 36 measured | $0.0461 | $0.0091 |
| Deck ratingrating, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 32 · 19 of 36 measured | 646 | 1,120 |
| Accurate decks% of tasks, higher is better 27 · 22 of 36 measured | 16.7% | 33.3% |
| ROASBench measured by Spring Prompt · 3 results | ||
| Overall scorescore out of 100, higher is better 35 · 34 of 36 measured | 17.5 | 24.1 |
| Audience scorescore out of 100, higher is better 35 · 30 of 36 measured | 30.9 | 48.4 |
| Business scorescore out of 100, higher is better 35 · 32 of 36 measured | 7.3 | 13.5 |
| SurveyBench measured by Spring Prompt · 3 results | ||
| Analysis ratingrating, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 26 · 9 of 27 measured | 561 | 1,233 |
| Cost per taskUS dollars, lower is better 10 · 4 of 27 measured | $0.0635 | $0.0185 |
| Analysis scorepoints of 100, higher is better 20 · 11 of 27 measured | 56.8 | 88.4 |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 122 · 20 of 249 measured | 16.9thinking reasoning | 43.4max reasoning |
| Humanity's Last Exam% of questions, higher is better 157 · 25 of 250 measured | 10.4%thinking reasoning | 44.4%max reasoning |
| Terminal-Bench 4.0% of tasks, higher is better 72 · 20 of 120 measured | 0.0%thinking reasoning | 32.8%max reasoning |
Show all 92 results (74 more)
| Measure | Claude Haiku 4.5 | Claude Haiku 5.5 |
|---|---|---|
| BulletBench measured by Spring Prompt · 7 results | ||
| Games lost on time · Blitz 3+2% of games, lower is better 1 · 11 of 23 measured | 0.0% | 37.5% |
| Games lost on time · Bullet 60s% of games, lower is better 31 · 36 of 62 measured | 25.0% | 37.5%no reasoning |
| Invalid moves · Blitz 3+2% of moves, lower is better 23 · 1 of 23 measured | 0.9% | 0.0% |
| Invalid moves · Bullet 60s% of moves, lower is better 56 · 1 of 62 measured | 0.3% | 0.0%no reasoning |
| Ladder Elo · Blitz 3+2ladder Elo, higher is better 20 · 14 of 23 measured | 491 | 557 |
| Median move time · Blitz 3+2milliseconds, lower is better 4 · 16 of 23 measured | 1.3 s | 6.2 s |
| Median move time · Bullet 60smilliseconds, lower is better 36 · 43 of 62 measured | 1.2 s | 1.7 sno reasoning |
| CatalogBench measured by Spring Prompt · 28 results | ||
| Channel compliance% of products, higher is better 1 · 1 of 36 measured | 100.0% | 100.0% |
| Channel rules broken% of products, lower is better 1 · 22 of 36 measured | 0.0% | 0.6% |
| Channel rules broken · sales brief% of products, lower is better 36 · 1 of 36 measured | 4.2% | 0.0% |
| Claims to check% of products, lower is better 34 · 25 of 34 measured | 22.6% | 0.6% |
| Claims to check · sales brief% of products, lower is better 34 · 24 of 34 measured | 44.6% | 1.2% |
| Conflicts caught% of conflicts, higher is better 1 · 1 of 36 measured | 100.0% | 100.0% |
| Content quality% of checks, higher is better 33 · 29 of 36 measured | 94.4% | 95.6% |
| Content quality · sales brief% of checks, higher is better 34 · 24 of 36 measured | 93.9% | 96.3% |
| Decision accuracy% of decisions, higher is better 35 · 28 of 36 measured | 90.6% | 95.6% |
| Failed outputs% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| Failed outputs · sales brief% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| Field accuracy% of missing fields, higher is better 18 · 34 of 36 measured | 93.7% | 91.0% |
| Invented values% of filled values, lower is better 36 · 11 of 36 measured | 6.2% | 0.8% |
| Missing UK information% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| Missing UK information · sales brief% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| Not findable% of products, lower is better 31 · 34 of 36 measured | 8.9% | 11.9% |
| Not findable · sales brief% of products, lower is better 27 · 34 of 36 measured | 6.6% | 10.7% |
| Publish-ready listings% of products, higher is better 33 · 26 of 36 measured | 12.5% | 47.0% |
| Publish-ready listings · sales brief% of products, higher is better 33 · 24 of 36 measured | 3.0% | 20.2% |
| Time per productseconds, lower is better 6 · 16 of 36 measured | 8.10 | 11.0 |
| Unsupported claimsclaims per product, lower is better 36 · 31 of 36 measured | 2.69 | 1.20 |
| Unsupported claims% of products, lower is better 36 · 28 of 36 measured | 73.2% | 32.1% |
| Unsupported claims · sales briefclaims per product, lower is better 36 · 25 of 36 measured | 5.32 | 2.46 |
| Unsupported claims · sales brief% of products, lower is better 36 · 25 of 36 measured | 91.7% | 70.2% |
| Wrong attributes% of products, lower is better 36 · 26 of 36 measured | 48.2% | 19.6% |
| Wrong attributes · sales brief% of products, lower is better 36 · 22 of 36 measured | 51.2% | 18.5% |
| Wrong category or variant% of products, lower is better 34 · 1 of 36 measured | 7.7% | 0.0% |
| Wrong category or variant · sales brief% of products, lower is better 34 · 1 of 36 measured | 6.6% | 0.0% |
| DeckBench measured by Spring Prompt · 14 results | ||
| Caveat dropped% of tasks, lower is better 1 · 16 of 36 measured | 0.0% | 16.7% |
| Clean layout% of tasks, higher is better 16 · 10 of 36 measured | 50.0% | 66.7% |
| Design quality% of the maximum, higher is better 32 · 22 of 36 measured | 47.6% | 61.1% |
| Draft figure quoted% of tasks, lower is better 29 · 21 of 36 measured | 50.0% | 33.3% |
| Findings missing% of tasks, lower is better 25 · 25 of 36 measured | 16.7% | 16.7% |
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 32 · 15 of 36 measured | 29.8% | 62.1% |
| Layout defects% of tasks, lower is better 16 · 10 of 36 measured | 50.0% | 33.3% |
| Misleading metric used% of tasks, lower is better 1 · 31 of 36 measured | 0.0% | 16.7% |
| Presentable decks% of tasks, higher is better 17 · 9 of 36 measured | 0.0% | 33.3% |
| Recommendation late or wrong% of tasks, lower is better 28 · 22 of 36 measured | 50.0% | 33.3% |
| Slides needing work% of tasks, lower is better 26 · 5 of 36 measured | 66.7% | 16.7% |
| Time per deckseconds, lower is better 6 · 7 of 36 measured | 51.7 | 62.1 |
| Unsupported claims% of tasks, lower is better 26 · 26 of 36 measured | 33.3% | 33.3% |
| Unsupported numbers% of tasks, lower is better 1 · 28 of 36 measured | 0.0% | 16.7% |
| ROASBench measured by Spring Prompt · 6 results | ||
| Consistency scorescore out of 100, higher is better 35 · 33 of 36 measured | 8.7 | 10.6 |
| Contribution profitsimulated US dollars, higher is better 36 · 32 of 36 measured | −$205,001 | $32,146 |
| Cost of a runUS dollars, lower is better 5 · 6 of 36 measured | $0.0298 | $0.0313 |
| Months over budgetmonths of 12, lower is better 34 · 1 of 36 measured | 2 | 0 |
| Planning scorescore out of 100, higher is better 27 · 21 of 36 measured | 54.4 | 54.8 |
| Return on ad spendprofit per $1 spent, higher is better 36 · 32 of 36 measured | −$0.21 | $0.03 |
| SurveyBench measured by Spring Prompt · 17 results | ||
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 26 · 8 of 27 measured | 24.7% | 59.9% |
| Leading or double-barrelled questions% of tasks, lower is better 24 · 12 of 27 measured | 66.7% | 33.3% |
| Mistaken next steps% of tasks, lower is better 22 · 13 of 27 measured | 55.6% | 22.2% |
| Numbers right% of answers, higher is better 26 · 1 of 27 measured | 63.7% | 100.0% |
| Objectives unanswerable% of tasks, lower is better 24 · 1 of 27 measured | 22.2% | 0.0% |
| Precision ignored% of tasks, lower is better 25 · 1 of 27 measured | 100.0% | 0.0% |
| Research scorepoints of 100, higher is better 20 · 10 of 27 measured | 64.9 | 90.2 |
| Sound research% of tasks, higher is better 17 · 13 of 27 measured | 0.0% | 11.1% |
| Stakeholder's view presumed% of tasks, lower is better 18 · 1 of 27 measured | 22.2% | 0.0% |
| Survey plan scorepoints of 100, higher is better 22 · 8 of 27 measured | 83.6 | 94.5 |
| Survey structure% of tasks, lower is better 1 · 12 of 27 measured | 0.0% | 22.2% |
| Time per taskseconds, lower is better 8 · 13 of 27 measured | 92.6 | 147 |
| Traps handled% of traps, higher is better 24 · 11 of 27 measured | 62.3% | 94.2% |
| Traps missed% of tasks, lower is better 19 · 4 of 27 measured | 88.9% | 22.2% |
| Unsupported findings% of tasks, lower is better 24 · 16 of 27 measured | 100.0% | 77.8% |
| Wrong numbers% of tasks, lower is better 25 · 1 of 27 measured | 100.0% | 0.0% |
| Wrong recommendation% of tasks, lower is better 19 · 5 of 27 measured | 44.4% | 11.1% |
| Artificial Analysis reported by Artificial Analysis · 2 results | ||
| Long-context reasoning (AA-LCR)% of questions, higher is better 89 · 21 of 241 measured | 74.3%thinking reasoning | 82.7%max reasoning |
| SciCode% of problems, higher is better 81 · 28 of 121 measured | 42.2%thinking reasoning | 55.0%max reasoning |
Compare them with other models
- Claude Haiku 4.5 vs Claude Opus 5.5Most intelligent available today
- Claude Haiku 4.5 vs GPT-6.1 TerraHighest-scoring overall at $1–3 per million tokens
- Claude Haiku 5.5 vs Claude Opus 5.5Most intelligent available today
- Claude Haiku 5.5 vs Muse EmberHighest-scoring overall at under $0.25 per million tokens
- Claude Haiku 5.5 vs Claude Fable 5.2Highest-scoring overall available today
Quick answers
Is Claude Haiku 4.5 better than Claude Haiku 5.5?
Claude Haiku 5.5 is the newer version and costs less, and it scores higher (Spring Prompt overall 38 against 7): upgrade.
Which is cheaper, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 is cheaper to run: $0.001 against $0.006 per product listing on CatalogBench (5.1 times less).
Which is better for product listings, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 ranks 23 of 35 models on sale for turning a sparse product feed and photos into listings that can go live, against 32 for Claude Haiku 4.5.
Which is better for decks from an analysis, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 ranks 19 of 35 models on sale for turning a finished analysis into a deck you could present as it is, against 31 for Claude Haiku 4.5.
Which is better for user surveys, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 ranks 9 of 26 models on sale for planning a user survey and reading its results without being misled, against 25 for Claude Haiku 4.5.
Which is better for marketing planning, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 ranks 33 of 35 models on sale for planning a year of ad spend without overspending, against 34 for Claude Haiku 4.5.
Which is better for agents and tool use, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5 ranks 28 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 98 for Claude Haiku 4.5.
Which has the bigger context window, Claude Haiku 4.5 or Claude Haiku 5.5?
Claude Haiku 5.5: 1,000,000 tokens, against 200,000 for Claude Haiku 4.5.
Run Claude Haiku 4.5 and Claude Haiku 5.5 on your own prompt.
Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.

