Compare / Claude Haiku 5.5 vs GPT-6 Luna
Claude Haiku 5.5 vs GPT-6 Luna
Which is cheaper, which is more capable, and which does better at each kind of work, from 94 results on 6 sources that measured both models.
Of the 94 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). See also Anthropic vs OpenAI, every model from both labs.
Claude Haiku 5.5 or GPT-6 Luna? The short answer
They score about the same, so pick GPT-6 Luna: it costs 2.5 times less per product listing on CatalogBench. Pick Claude Haiku 5.5 if speed matters: it writes faster.
- Spring Prompt overall: Claude Haiku 5.5 38, GPT-6 Luna 38 (out of 100; how the score works).
- GPT-6 Luna is cheaper to run: $0.000 against $0.001 per product listing on CatalogBench (2.5 times less).
- Claude Haiku 5.5 scores higher on the Artificial Analysis Intelligence Index: 43.4 against 38.1.
- Claude Haiku 5.5 writes faster: 240 tokens a second against 127.
- Claude Haiku 5.5 is newer: released 7 Oct 2026, against 22 Sep 2026 for GPT-6 Luna.
- On our own benchmarks, on the results both models have: Claude Haiku 5.5 does better for user surveys; GPT-6 Luna for product listings, decks from an analysis 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 5.5 | GPT-6 Luna | |
|---|---|---|
| Price per million tokens | $0.10 in · $0.50 out | $0.10 in · $0.50 out |
| Blended (3 in : 1 out) | $0.20 | $0.20 |
| Cheapest host, blended | $0.20 (Amazon Bedrock) | $0.20 (Amazon Bedrock) |
| Cost per product listing (CatalogBench) | $0.001 | $0.000 |
| Cached input, per million | – | $0.010 cached input |
| Intelligence Index | 43.4 | 38.1 |
| Spring Prompt overall | 38 | 38 |
| Output speed, tokens a second | 240 | 127 |
| Context window | 1,000,000 tokens | 1,050,000 tokens |
| Released | 7 Oct 2026 (newer) | 22 Sep 2026 |
| 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
- AI Checkpoint: Sol’s new fast lane 9 Oct 2026
- 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
- 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 5.5 | GPT-6 Luna |
|---|---|---|
| BulletBench measured by Spring Prompt · 3 results | ||
| Cost per game · Blitz 3+2US dollars, lower is better 4 · 1 of 23 measured | $0.0164 | $0.0072 |
| Cost per game · Bullet 60sUS dollars, lower is better 23 · 21 of 62 measured | $0.0019no reasoning | $0.0016no reasoning |
| Ladder Elo · Bullet 60sladder Elo, higher is better 50 · 52 of 62 measured | 310no reasoning | 249no reasoning |
| CatalogBench measured by Spring Prompt · 3 results | ||
| Cost per productUS dollars, lower is better 4 · 2 of 36 measured | $0.0012 | $0.0005 |
| Reliably publish-ready% of products, higher is better 26 · 13 of 35 measured | 26.8% | 55.4% |
| Reliably publish-ready · sales brief% of products, higher is better 20 · 7 of 35 measured | 10.7% | 51.8% |
| DeckBench measured by Spring Prompt · 3 results | ||
| Cost per deckUS dollars, lower is better 4 · 3 of 36 measured | $0.0091 | $0.0087 |
| 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. 19 · 12 of 36 measured | 1,120 | 1,252 |
| Accurate decks% of tasks, higher is better 22 · 17 of 36 measured | 33.3% | 50.0% |
| ROASBench measured by Spring Prompt · 3 results | ||
| Overall scorescore out of 100, higher is better 34 · 7 of 36 measured | 24.1 | 53.4 |
| Audience scorescore out of 100, higher is better 30 · 7 of 36 measured | 48.4 | 68.3 |
| Business scorescore out of 100, higher is better 32 · 7 of 36 measured | 13.5 | 54.8 |
| 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. 9 · 14 of 27 measured | 1,233 | 1,072 |
| Cost per taskUS dollars, lower is better 4 · 1 of 27 measured | $0.0185 | $0.0088 |
| Analysis scorepoints of 100, higher is better 11 · 7 of 27 measured | 88.4 | 90.8 |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 20 · 39 of 249 measured | 43.4max reasoning | 38.1max reasoning |
| Humanity's Last Exam% of questions, higher is better 25 · 53 of 250 measured | 44.4%max reasoning | 38.5%max reasoning |
| Terminal-Bench 4.0% of tasks, higher is better 20 · 37 of 120 measured | 32.8%max reasoning | 12.6%max reasoning |
Show all 94 results (76 more)
| Measure | Claude Haiku 5.5 | GPT-6 Luna |
|---|---|---|
| BulletBench measured by Spring Prompt · 7 results | ||
| Games lost on time · Blitz 3+2% of games, lower is better 11 · 15 of 23 measured | 37.5% | 50.0% |
| Games lost on time · Bullet 60s% of games, lower is better 36 · 39 of 62 measured | 37.5%no reasoning | 50.0%no reasoning |
| Invalid moves · Blitz 3+2% of moves, lower is better 1 · 1 of 23 measured | 0.0% | 0.0% |
| Invalid moves · Bullet 60s% of moves, lower is better 1 · 1 of 62 measured | 0.0%no reasoning | 0.0%no reasoning |
| Ladder Elo · Blitz 3+2ladder Elo, higher is better 14 · 14 of 23 measured | 557 | 557 |
| Median move time · Blitz 3+2milliseconds, lower is better 16 · 19 of 23 measured | 6.2 s | 7.0 s |
| Median move time · Bullet 60smilliseconds, lower is better 43 · 39 of 62 measured | 1.7 sno reasoning | 1.5 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 22 · 1 of 36 measured | 0.6% | 0.0% |
| Channel rules broken · sales brief% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| Claims to check% of products, lower is better 25 · 1 of 34 measured | 0.6% | 0.0% |
| Claims to check · sales brief% of products, lower is better 24 · 1 of 34 measured | 1.2% | 0.0% |
| Conflicts caught% of conflicts, higher is better 1 · 1 of 36 measured | 100.0% | 100.0% |
| Content quality% of checks, higher is better 29 · 20 of 36 measured | 95.6% | 97.0% |
| Content quality · sales brief% of checks, higher is better 24 · 18 of 36 measured | 96.3% | 97.4% |
| Decision accuracy% of decisions, higher is better 28 · 16 of 36 measured | 95.6% | 97.0% |
| 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 34 · 18 of 36 measured | 91.0% | 93.7% |
| Invented values% of filled values, lower is better 11 · 18 of 36 measured | 0.8% | 1.7% |
| 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 34 · 32 of 36 measured | 11.9% | 10.1% |
| Not findable · sales brief% of products, lower is better 34 · 32 of 36 measured | 10.7% | 10.1% |
| Publish-ready listings% of products, higher is better 26 · 17 of 36 measured | 47.0% | 67.9% |
| Publish-ready listings · sales brief% of products, higher is better 24 · 8 of 36 measured | 20.2% | 67.9% |
| Time per productseconds, lower is better 16 · 8 of 36 measured | 11.0 | 8.20 |
| Unsupported claimsclaims per product, lower is better 31 · 13 of 36 measured | 1.20 | 0.31 |
| Unsupported claims% of products, lower is better 28 · 13 of 36 measured | 32.1% | 8.9% |
| Unsupported claims · sales briefclaims per product, lower is better 25 · 8 of 36 measured | 2.46 | 0.35 |
| Unsupported claims · sales brief% of products, lower is better 25 · 8 of 36 measured | 70.2% | 8.3% |
| Wrong attributes% of products, lower is better 26 · 24 of 36 measured | 19.6% | 18.5% |
| Wrong attributes · sales brief% of products, lower is better 22 · 23 of 36 measured | 18.5% | 19.1% |
| Wrong category or variant% of products, lower is better 1 · 7 of 36 measured | 0.0% | 1.8% |
| Wrong category or variant · sales brief% of products, lower is better 1 · 1 of 36 measured | 0.0% | 0.0% |
| DeckBench measured by Spring Prompt · 14 results | ||
| Caveat dropped% of tasks, lower is better 16 · 1 of 36 measured | 16.7% | 0.0% |
| Clean layout% of tasks, higher is better 10 · 10 of 36 measured | 66.7% | 66.7% |
| Design quality% of the maximum, higher is better 22 · 10 of 36 measured | 61.1% | 78.1% |
| Draft figure quoted% of tasks, lower is better 21 · 21 of 36 measured | 33.3% | 33.3% |
| Findings missing% of tasks, lower is better 25 · 1 of 36 measured | 16.7% | 0.0% |
| 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. 15 · 9 of 36 measured | 62.1% | 70.9% |
| Layout defects% of tasks, lower is better 10 · 10 of 36 measured | 33.3% | 33.3% |
| Misleading metric used% of tasks, lower is better 31 · 1 of 36 measured | 16.7% | 0.0% |
| Presentable decks% of tasks, higher is better 9 · 17 of 36 measured | 33.3% | 0.0% |
| Recommendation late or wrong% of tasks, lower is better 22 · 17 of 36 measured | 33.3% | 16.7% |
| Slides needing work% of tasks, lower is better 5 · 5 of 36 measured | 16.7% | 16.7% |
| Time per deckseconds, lower is better 7 · 17 of 36 measured | 62.1 | 101 |
| Unsupported claims% of tasks, lower is better 26 · 1 of 36 measured | 33.3% | 0.0% |
| Unsupported numbers% of tasks, lower is better 28 · 1 of 36 measured | 16.7% | 0.0% |
| ROASBench measured by Spring Prompt · 6 results | ||
| Consistency scorescore out of 100, higher is better 33 · 10 of 36 measured | 10.6 | 35.8 |
| Contribution profitsimulated US dollars, higher is better 32 · 9 of 36 measured | $32,146 | $738,879 |
| Cost of a runUS dollars, lower is better 6 · 2 of 36 measured | $0.0313 | $0.0142 |
| Months over budgetmonths of 12, lower is better 1 · 1 of 36 measured | 0 | 0 |
| Planning scorescore out of 100, higher is better 21 · 20 of 36 measured | 54.8 | 54.8 |
| Return on ad spendprofit per $1 spent, higher is better 32 · 10 of 36 measured | $0.03 | $0.66 |
| 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. 8 · 17 of 27 measured | 59.9% | 47.1% |
| Leading or double-barrelled questions% of tasks, lower is better 12 · 5 of 27 measured | 33.3% | 11.1% |
| Mistaken next steps% of tasks, lower is better 13 · 5 of 27 measured | 22.2% | 11.1% |
| Numbers right% of answers, higher is better 1 · 21 of 27 measured | 100.0% | 98.0% |
| Objectives unanswerable% of tasks, lower is better 1 · 1 of 27 measured | 0.0% | 0.0% |
| Precision ignored% of tasks, lower is better 1 · 1 of 27 measured | 0.0% | 0.0% |
| Research scorepoints of 100, higher is better 10 · 7 of 27 measured | 90.2 | 91.7 |
| Sound research% of tasks, higher is better 13 · 6 of 27 measured | 11.1% | 33.3% |
| Stakeholder's view presumed% of tasks, lower is better 1 · 12 of 27 measured | 0.0% | 11.1% |
| Survey plan scorepoints of 100, higher is better 8 · 9 of 27 measured | 94.5 | 94.0 |
| Survey structure% of tasks, lower is better 12 · 12 of 27 measured | 22.2% | 22.2% |
| Time per taskseconds, lower is better 13 · 12 of 27 measured | 147 | 138 |
| Traps handled% of traps, higher is better 11 · 11 of 27 measured | 94.2% | 94.2% |
| Traps missed% of tasks, lower is better 4 · 8 of 27 measured | 22.2% | 33.3% |
| Unsupported findings% of tasks, lower is better 16 · 7 of 27 measured | 77.8% | 22.2% |
| Wrong numbers% of tasks, lower is better 1 · 14 of 27 measured | 0.0% | 11.1% |
| Wrong recommendation% of tasks, lower is better 5 · 5 of 27 measured | 11.1% | 11.1% |
| Artificial Analysis reported by Artificial Analysis · 4 results | ||
| Long-context reasoning (AA-LCR)% of questions, higher is better 21 · 13 of 241 measured | 82.7%max reasoning | 83.3%max reasoning |
| Output speedtokens per second, higher is better 10 · 37 of 78 measured | 240max reasoning | 127max reasoning |
| SciCode% of problems, higher is better 28 · 31 of 121 measured | 55.0%max reasoning | 54.6%max reasoning |
| Time to first answer tokenseconds, lower is better 40 · 20 of 78 measured | 6.41low reasoning | 0.74no reasoning |
Compare them with other models
- Anthropic vs OpenAIEvery model from both labs, best against best
- 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
- GPT-6 Luna vs Claude Opus 5.5Most intelligent available today
Quick answers
Is Claude Haiku 5.5 better than GPT-6 Luna?
They score about the same, so pick GPT-6 Luna: it costs 2.5 times less per product listing on CatalogBench. Pick Claude Haiku 5.5 if speed matters: it writes faster.
Which is cheaper, Claude Haiku 5.5 or GPT-6 Luna?
GPT-6 Luna is cheaper to run: $0.000 against $0.001 per product listing on CatalogBench (2.5 times less).
Which is faster, Claude Haiku 5.5 or GPT-6 Luna?
Claude Haiku 5.5 writes about 240 tokens a second on its usual API, against 127 for GPT-6 Luna (Artificial Analysis).
Which is better for product listings, Claude Haiku 5.5 or GPT-6 Luna?
GPT-6 Luna ranks 10 of 35 models on sale for turning a sparse product feed and photos into listings that can go live, against 23 for Claude Haiku 5.5.
Which is better for decks from an analysis, Claude Haiku 5.5 or GPT-6 Luna?
GPT-6 Luna ranks 12 of 35 models on sale for turning a finished analysis into a deck you could present as it is, against 19 for Claude Haiku 5.5.
Which is better for user surveys, Claude Haiku 5.5 or GPT-6 Luna?
Claude Haiku 5.5 ranks 9 of 26 models on sale for planning a user survey and reading its results without being misled, against 14 for GPT-6 Luna.
Which is better for marketing planning, Claude Haiku 5.5 or GPT-6 Luna?
GPT-6 Luna ranks 7 of 35 models on sale for planning a year of ad spend without overspending, against 33 for Claude Haiku 5.5.
Which has the bigger context window, Claude Haiku 5.5 or GPT-6 Luna?
GPT-6 Luna: 1,050,000 tokens, against 1,000,000 for Claude Haiku 5.5.
Run Claude Haiku 5.5 and GPT-6 Luna 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.

