Coding work
Code generation, repair, and software-engineering tasks
GPT-5.6 Luna
GPT-5.6 Luna leads on the public coding lane, 64.5 to 19.8, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
Claude Haiku 4.5 vs GPT-5.6 Luna
Updated September 24, 2026. Rank says GPT-5.6 Luna is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Decision reading
GPT-5.6 Luna has the higher public score, 65.6 versus 42.48, and the 90% score intervals do not overlap. 7 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 7
- Claude Haiku 4.5 only
- 3
- GPT-5.6 Luna only
- 22
- Like-for-like categories
- 2 / 8
Estimated: Claude Haiku 4.5 · Supported: GPT-5.6 LunaHow the comparison works
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Agentic work
Tool use, computer use, and multi-step task completion
GPT-5.6 Luna
GPT-5.6 Luna leads on the public agentic lane, 55.3 to 21.7, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: strongerLong documents
Prompts that approach the documented context limit
GPT-5.6 Luna
GPT-5.6 Luna has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Chat turn cost
1K fresh input + 500 output tokens
GPT-5.6 Luna
GPT-5.6 Luna has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesRepository review cost
50K fresh input + 3K output tokens
GPT-5.6 Luna
GPT-5.6 Luna has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Claude Haiku 4.5 does not fit this workload in one request.
Confidence: listed-rates
Which one for a specific job
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
19.8Claude Haiku 4.564.5GPT-5.6 Luna
Like-for-like · BenchAlign v5.7
GPT-5.6 Luna leads the like-for-like coding row.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
3 Claude Haiku 4.5 only
7 Shared
22 GPT-5.6 Luna only
1 category rests on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
FrontierMath v2 (Tiers 1-3)Math
Normalized gap 72.7Claude Haiku 4.5:5.903%GPT-5.6 Luna:78.600%FrontierMath v2 (Tier 4)Math
Normalized gap 56.4Claude Haiku 4.5:2.083%GPT-5.6 Luna:58.500%MMLU-Pro (Vals)Knowledge
Normalized gap 7.3Claude Haiku 4.5:78.7%GPT-5.6 Luna:86.0%
Bars run 0–100 on each benchmark’s normalized display scale
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign v5.7 lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
Agentic
Like-for-like- Claude Haiku 4.5
- 21.7
- Supported · #85/105
- GPT-5.6 Luna
- 55.3
- Supported · #28/105
- Basis
- BenchAlign v5.7 lane · 2 vs 9 public rows
- Reading
- GPT-5.6 Luna leads
Coding
Like-for-like- Claude Haiku 4.5
- 19.8
- Supported · #117/135
- GPT-5.6 Luna
- 64.5
- Supported · #9/135
- Basis
- BenchAlign v5.7 lane · 4 vs 7 public rows
- Reading
- GPT-5.6 Luna leads
Knowledge
Directional only- Claude Haiku 4.5
- 34.5
- Estimated · #111/158
- GPT-5.6 Luna
- 64.6
- Supported · #22/158
- Basis
- BenchAlign v5.7 lane · 2 vs 6 public rows
- Reading
- Directional only
Reasoning
Not comparable- Claude Haiku 4.5
- Not ranked
- GPT-5.6 Luna
- 54.7
- #18/19
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Claude Haiku 4.5
- Not ranked
- GPT-5.6 Luna
- 67.1
- #22/50
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Haiku 4.5
- Not ranked
- GPT-5.6 Luna
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Haiku 4.5
- Not ranked
- GPT-5.6 Luna
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- Claude Haiku 4.5
- 28.8
- Unranked · 2 rankable rows
- GPT-5.6 Luna
- 94.2
- Unranked · 3 rankable rows
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Not comparable
| Category | Claude Haiku 4.5 | GPT-5.6 Luna | Basis | Reading |
|---|---|---|---|---|
| Agentic | 21.7Supported · #85/105 | 55.3Supported · #28/105 | Like-for-likeBenchAlign v5.7 lane · 2 vs 9 public rows | GPT-5.6 Luna leads |
| Coding | 19.8Supported · #117/135 | 64.5Supported · #9/135 | Like-for-likeBenchAlign v5.7 lane · 4 vs 7 public rows | GPT-5.6 Luna leads |
| Knowledge | 34.5Estimated · #111/158 | 64.6Supported · #22/158 | Directional onlyBenchAlign v5.7 lane · 2 vs 6 public rows | Directional only |
| Reasoning | Not ranked | 54.7#18/19 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Multimodal | Not ranked | 67.1#22/50 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 28.8Unranked · 2 rankable rows | 94.2Unranked · 3 rankable rows | Not comparableProvisional lane · 2 vs 2 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.7) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Supported evidence per lane · bars run 0–100Methodology
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- Claude Haiku 4.5
- $0.0035
- Fits in one request
- GPT-5.6 Luna
- $0.0008
- Fits in one request
GPT-5.6 Luna has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Haiku 4.5
- $0.065
- Fits in one request
- GPT-5.6 Luna
- $0.0136
- Fits in one request
GPT-5.6 Luna has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Claude Haiku 4.5
- $0.09
- Does not fit in one request
- GPT-5.6 Luna
- $0.02
- Fits in one request
Claude Haiku 4.5 does not fit this workload in one request.
Cached input falls back to the list input rate only where a cached rate is unpublished
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Haiku 4.5GPT-5.6 Luna
Context window
Maximum documented context; output-token limits may be lower.
Claude Haiku 4.5
GPT-5.6 Luna
1.05M
OpenAI model catalogAPI model ID
Claude Haiku 4.5
claude-haiku-4-5-20251001
Claude API pricingGPT-5.6 Luna
gpt-5.6-luna
OpenAI model catalogCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Haiku 4.5
$0.1 per 1M cached input tokens
Claude API pricingGPT-5.6 Luna
$0.02 per 1M cached input tokens
OpenAI pricingDocumented inputs
Claude Haiku 4.5
Not sourced
GPT-5.6 Luna
text, image
OpenAI model catalogDocumented outputs
Claude Haiku 4.5
Not sourced
GPT-5.6 Luna
Provider availability
Claude Haiku 4.5
Not sourced
GPT-5.6 Luna
Generally Available · OpenAI Responses API
OpenAI model catalogReasoning profile
Claude Haiku 4.5
Non-Reasoning
GPT-5.6 Luna
Reasoning
Weight access
Claude Haiku 4.5
Proprietary
GPT-5.6 Luna
Proprietary
License
Claude Haiku 4.5
Proprietary
GPT-5.6 Luna
Proprietary
Release date
Claude Haiku 4.5
2025-10-15
GPT-5.6 Luna
2026-07-09
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- GPT-5.6 Luna has the higher public score, 65.6 versus 42.48, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.065 vs $0.0136. Cache-heavy agent loop: $0.09 vs $0.02.
- Context tradeoff
- GPT-5.6 Luna has the larger documented window (1.05M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Claude Haiku 4.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the higher public score, 65.6 versus 42.48, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.
Which is better for coding, Claude Haiku 4.5 or GPT-5.6 Luna?
GPT-5.6 Luna leads the public coding lane, 64.5 to 19.8, with Supported evidence for both models and non-overlapping 90% intervals.
Which is better for agentic tasks, Claude Haiku 4.5 or GPT-5.6 Luna?
GPT-5.6 Luna leads the public agentic tasks lane, 55.3 to 21.7, with Supported evidence for both models and non-overlapping 90% intervals.
Which costs less, Claude Haiku 4.5 or GPT-5.6 Luna?
For the stated presets, chat costs $0.0035 on Claude Haiku 4.5 and $0.0008 on GPT-5.6 Luna; repository review costs $0.065 and $0.0136; the cache-heavy agent loop costs $0.09 and $0.02. Claude Haiku 4.5 does not fit this workload in one request.
Which has the larger context window, Claude Haiku 4.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the larger documented context window: 1.05M, compared with 200K.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence32 rows
Agentic
JobBench
Claude Haiku 4.516.0%SourceGPT-5.6 Luna—Not directly comparable
Terminal-Bench 2.1 (Vals)
Claude Haiku 4.543.8%SourceGPT-5.6 Luna79.0%SourceGPT-5.6 Luna leads this result
Terminal-Bench 3.0
Claude Haiku 4.5—GPT-5.6 Luna14.3%SourceNot directly comparable
Terminal-Bench 2.1
Claude Haiku 4.5—GPT-5.6 Luna84.7%SourceNot directly comparable
BrowseComp
Claude Haiku 4.5—GPT-5.6 Luna83.3%SourceNot directly comparable
OSWorld 2.0
Claude Haiku 4.5—GPT-5.6 Luna45.6%SourceNot directly comparable
CyberGym
Claude Haiku 4.5—GPT-5.6 Luna77.9%SourceNot directly comparable
ExploitGym
Claude Haiku 4.5—GPT-5.6 Luna12.4%SourceNot directly comparable
Toolathlon
Claude Haiku 4.5—GPT-5.6 Luna53.4%SourceNot directly comparable
ApprenticeBench
Claude Haiku 4.5—GPT-5.6 Luna7%SourceNot directly comparable
Coding
SWE-bench Verified
Claude Haiku 4.573.3%SourceGPT-5.6 Luna—Not directly comparable
VulcanBench v3
Claude Haiku 4.576.2%SourceGPT-5.6 Luna85.5%SourceGPT-5.6 Luna leads this result
LiveCodeBench (Vals)
Claude Haiku 4.541.2%SourceGPT-5.6 Luna—Not directly comparable
SWE-bench (Vals)
Claude Haiku 4.566.6%SourceGPT-5.6 Luna93.0%SourceGPT-5.6 Luna leads this result
SWE-bench Pro
Claude Haiku 4.5—GPT-5.6 Luna62.7%SourceNot directly comparable
Terminal-Bench 2.1
Claude Haiku 4.5—GPT-5.6 Luna84.7%SourceNot directly comparable
DeepSWE
Claude Haiku 4.5—GPT-5.6 Luna67.2%SourceNot directly comparable
FrontierCode 1.1 Extended
Claude Haiku 4.5—GPT-5.6 Luna55.1%SourceNot directly comparable
cursorBench32
Claude Haiku 4.5—GPT-5.6 Luna61.1%SourceNot directly comparable
Reasoning
Multimodal
Knowledge
GPQA Diamond (Vals)
Claude Haiku 4.572.2%SourceGPT-5.6 Luna91.7%SourceGPT-5.6 Luna leads this result
MMLU-Pro (Vals)
Claude Haiku 4.578.7%SourceGPT-5.6 Luna86.0%SourceGPT-5.6 Luna leads this result
GPQA
Claude Haiku 4.5—GPT-5.6 Luna92.3%SourceNot directly comparable
GPQA-D
Claude Haiku 4.5—GPT-5.6 Luna92.3%SourceNot directly comparable
HealthBench Professional
Claude Haiku 4.5—GPT-5.6 Luna55.7%SourceNot directly comparable
HealthBench Hard
Claude Haiku 4.5—GPT-5.6 Luna32.0%SourceNot directly comparable
Math
FrontierMath v2 (Tiers 1-3)
Claude Haiku 4.55.903%SourceGPT-5.6 Luna78.600%SourceGPT-5.6 Luna leads this result
FrontierMath v2 (Tier 4)
Claude Haiku 4.52.083%SourceGPT-5.6 Luna58.500%SourceGPT-5.6 Luna leads this result
FrontierMath (legacy)
Claude Haiku 4.5—GPT-5.6 Luna78.6%SourceNot directly comparable
32 public results · 7 shared
Watch Claude Haiku 4.5 vs GPT-5.6 Luna
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.
Last updated September 24, 2026