Updated October 17, 2026. Rank says GPT-6 Astra is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
GPT-6 Astra has the higher public score estimate, 83.88 versus 79.63, but the 90% score intervals overlap. Treat that as a lead, not a settled winner. 24 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
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.
Code generation, repair, and software-engineering tasks
GPT-6 Astra
GPT-6 Astra has the higher public coding point estimate, 75.5 to 70.7, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Tool use, computer use, and multi-step task completion
Claude Opus 5
Claude Opus 5 has the higher public agentic point estimate, 78.5 to 74.8, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Prompts that approach the documented context limit
GPT-6 Astra
GPT-6 Astra has the larger documented context window.
1K fresh input + 500 output tokens
Claude Opus 5
Claude Opus 5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
200K cached + 20K fresh input + 10K output tokens
Claude Opus 5
Claude Opus 5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
50K fresh input + 3K output tokens
Claude Opus 5
Claude Opus 5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
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.
Like-for-like · BenchAlign v5.8
GPT-6 Astra has the higher coding point estimate. Conditional score ranges do not establish rank confidence.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
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.
ARC-AGI-3Reasoning
Normalized gap 32.6FrontierMath v2 (Tier 4)Math
Normalized gap 24.4AutomationBenchAgentic
Normalized gap 15.4FrontierMath v2 (Tiers 1-3)Math
Normalized gap 8.1ARC-AGI-2Reasoning
Normalized gap 4.6Each row shows the public-lane category score for both models: the BenchAlign v5.8 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.
| Category | Claude Opus 5 | GPT-6 Astra | Basis | Reading |
|---|---|---|---|---|
| Agentic | 78.5Supported · #3/128 | 74.8Supported · #7/128 | Like-for-likeBenchAlign v5.8 lane · 21 vs 14 public rows | Claude Opus 5 leads · intervals overlap |
| Coding | 70.7Supported · #9/150 | 75.5Supported · #4/150 | Like-for-likeBenchAlign v5.8 lane · 18 vs 6 public rows | GPT-6 Astra leads · intervals overlap |
| Reasoning | 81.0#11/28 | 93.3#1/28 | Like-for-likeProvisional lane · 2 vs 2 weighted rows | GPT-6 Astra leads |
| Knowledge | 82.0Supported · #7/184 | 86.2Supported · #3/184 | Like-for-likeBenchAlign v5.8 lane · 19 vs 7 public rows | GPT-6 Astra leads · intervals overlap |
| Multimodal | 92.2#2/54 | 87.3Unranked · 3 rankable rows | Not comparableProvisional lane · 1 vs 0 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 | 91.6Unranked · 7 rankable rows | 97.3Unranked · 2 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 v5.8) 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.
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.
1K fresh input + 500 output tokens
Claude Opus 5 has the lower modeled cost
Costs use the listed standard API rates.
50K fresh input + 3K output tokens
Claude Opus 5 has the lower modeled cost
Costs use the listed standard API rates.
200K cached + 20K fresh input + 10K output tokens
Claude Opus 5 has the lower modeled cost
Costs use the listed standard API rates.
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Claude Opus 5
GPT-6 Astra
Claude Opus 5
claude-opus-5
Anthropic model overviewGPT-6 Astra
gpt-6-astra
OpenAI GPT-6 Astra model documentationA missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Opus 5
$0.5 per 1M cached input tokens
Claude API pricingGPT-6 Astra
$1 per 1M cached input tokens
OpenAI GPT-6 Astra model documentationClaude Opus 5
text, image
Anthropic model overviewGPT-6 Astra
text, image
OpenAI GPT-6 Astra model documentationClaude Opus 5
GPT-6 Astra
Claude Opus 5
Generally Available · Claude API
Anthropic model overviewGPT-6 Astra
Limited Availability · OpenAI Responses API, OpenAI Chat Completions API, ChatGPT Plus, Pro, Business, and Enterprise, Microsoft Azure, AWS Bedrock
OpenAI GPT-6 Astra model documentationClaude Opus 5
Reasoning
GPT-6 Astra
Reasoning
Claude Opus 5
Proprietary
GPT-6 Astra
Proprietary
Claude Opus 5
Proprietary
GPT-6 Astra
Proprietary
Claude Opus 5
2026-07-24
GPT-6 Astra
2026-09-03
GPT-6 Astra has the higher public score estimate, 83.88 versus 79.63, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
GPT-6 Astra has the higher public coding point estimate, 75.5 to 70.7, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Claude Opus 5 has the higher public agentic tasks point estimate, 78.5 to 74.8, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
For the stated presets, chat costs $0.0175 on Claude Opus 5 and $0.035 on GPT-6 Astra; repository review costs $0.325 and $0.65; the cache-heavy agent loop costs $0.45 and $0.9. Costs use the listed standard API rates.
GPT-6 Astra has the larger documented context window: 1.05M, compared with 1M.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Terminal-Bench 3.0
Not directly comparable
BrowseComp
GPT-6 Astra leads this result
HLE w/ tools
Claude Opus 5 leads this result
DeepSearchQA
Not directly comparable
DRACO
Not directly comparable
BrowseComp (10-agent, prerelease)
Not directly comparable
OSWorld 2.0
GPT-6 Astra leads this result
MCP Atlas
Not directly comparable
MCP-Atlas claim coverage
Not directly comparable
LAB all-pass (Anthropic harness)
Not directly comparable
LAB criterion-pass (Anthropic harness)
Not directly comparable
LAB all-pass (Harvey held-out)
Not directly comparable
LAB criterion-pass (Harvey held-out)
Not directly comparable
Toolathlon-Verified
Not directly comparable
Toolathlon Verified Pass@3
Not directly comparable
Toolathlon Verified Pass³
Not directly comparable
Toolathlon Verified avg. turns
Not directly comparable
AutomationBench
GPT-6 Astra leads this result
Terminal-Bench 2.1 (Vals)
GPT-6 Astra leads this result
ApprenticeBench
Shared sourceGPT-6 Astra leads this result
MazeBench
GPT-6 Astra leads this result
Terminal-Bench 4.0
Not directly comparable
Terminal-Bench-Science 0.1
Not directly comparable
ExploitGym
Not directly comparable
Agents' Last Exam
Not directly comparable
CWE-bench v1
Not directly comparable
Hermes Index
Not directly comparable
Hermes Bench
Not directly comparable
Bug Hunt Bench
Shared sourceGPT-6 Astra leads this result
SWE-bench Verified
Not directly comparable
SWE-bench Pro
Not directly comparable
SWE Multilingual
Not directly comparable
SWE Multimodal
Not directly comparable
DeepSWE
GPT-6 Astra leads this result
FrontierCode 1.1 Main
Claude Opus 5 leads this result
FrontierCode 1.1 Extended
GPT-6 Astra leads this result
FrontierSWE v2
Shared sourceGPT-6 Astra leads this result
ProgramBench (episode 1)
Not directly comparable
ProgramBench
Not directly comparable
CursorBench 3.2
Not directly comparable
VulcanBench v3
Not directly comparable
VulcanBench CII v1
Not directly comparable
LiveCodeBench (Vals)
Not directly comparable
SWE-bench (Vals)
Not directly comparable
CursorBench 4.0
Not directly comparable
PostTrainBench v1.1
GPT-6 Astra leads this result
ARC-AGI-1
GPT-6 Astra leads this result
ARC-AGI-2
GPT-6 Astra leads this result
ARC-AGI-3
GPT-6 Astra leads this result
STEMBench
Not directly comparable
GeneBench-Pro
Not directly comparable
MRCR v2 256K-512K
Not directly comparable
MRCR v2 512K-1M
Not directly comparable
GraphWalks BFS 256K–1M
Not directly comparable
Chartography (no tools)
Not directly comparable
Chartography (tools)
Not directly comparable
BenchCAD Vision2Code (no tools)
Not directly comparable
BenchCAD Vision2Code (tools)
GPT-6 Astra leads this result
GDP.pdf (no tools)
Not directly comparable
GDP.pdf (tools)
Not directly comparable
OfficeQA
Not directly comparable
OfficeQA Pro
Not directly comparable
ScreenSpot Pro
Not directly comparable
HLE
Not directly comparable
HLE w/o tools
Not directly comparable
HLE-Verified
Not directly comparable
LABBench2
Not directly comparable
HealthBench (raw)
Claude Opus 5 leads this result
HealthBench (length-adjusted)
GPT-6 Astra leads this result
HealthBench Professional
GPT-6 Astra leads this result
HealthBench Professional (raw)
Claude Opus 5 leads this result
BioMysteryBench (human-solvable)
Not directly comparable
BioMysteryBench (human-difficult)
Not directly comparable
SpatialBench Verified
Not directly comparable
SingleCellBench
Not directly comparable
ProteinGym Hard
Not directly comparable
Protein Design
Not directly comparable
Organic chemistry V2
Not directly comparable
Protocols (troubleshooting)
Not directly comparable
Protocols (understanding)
Not directly comparable
GPQA Diamond (Vals)
Not directly comparable
MMLU-Pro (Vals)
Not directly comparable
GPQA
Not directly comparable
GPQA-D
Not directly comparable
HealthBench Hard
Not directly comparable
Gray Swan IPI (15 attempts)
Shared sourceClaude Opus 5 leads this result
IMO 2026
Not directly comparable
RiemannBench (no tools)
Not directly comparable
RiemannBench (tools)
Not directly comparable
ArXivMath Jun. 2026 (no tools)
Not directly comparable
ArXivMath Jun. 2026 (tools)
Not directly comparable
FrontierMath v2 (Tiers 1-3)
Shared sourceGPT-6 Astra leads this result
FrontierMath v2 (Tier 4)
GPT-6 Astra leads this result
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 5,600+ readers.
Last updated October 17, 2026
