Agentic
Directional only- Atria Dawn Preview
- 57.9
- Estimated · #31/153
- Claude Opus 4.8
- 62.5
- Supported · #14/153
- Basis
- BenchAlign lane · 11 vs 12 public rows
- Reading
- Directional only
Model A
Shanghai Artificial Intelligence Laboratory
—
Evidence status unavailable
90% interval unavailable
Atria Dawn Preview vs Claude Opus 4.8
Updated September 14, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
3 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
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.
Long documents
Prompts that approach the documented context limit
Claude Opus 4.8
Claude Opus 4.8 has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Atria Dawn Preview is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Atria Dawn Preview is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: rate-fallback
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
10 Atria Dawn Preview only
3 Shared
35 Claude Opus 4.8 only
- Shared results
- 3
- Atria Dawn Preview only
- 10
- Claude Opus 4.8 only
- 35
- Like-for-like categories
- 0 / 8
2 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign 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.
Coding
Directional only- Atria Dawn Preview
- 51.6
- Estimated · #51/152
- Claude Opus 4.8
- 66.5
- Supported · #12/152
- Basis
- BenchAlign lane · 2 vs 10 public rows
- Reading
- Directional only
Reasoning
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- 55.5
- #18/20
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- 71.4
- Supported · #9/183
- Basis
- BenchAlign lane · 0 vs 6 public rows
- Reading
- Not comparable
Math
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- 65.4
- #2/7
- Basis
- Provisional lane · 0 vs 3 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- 87.8
- #3/48
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Atria Dawn Preview
- Not ranked
- Claude Opus 4.8
- 75.4
- #61/123
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | Atria Dawn Preview | Claude Opus 4.8 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 57.9Estimated · #31/153 | 62.5Supported · #14/153 | Directional onlyBenchAlign lane · 11 vs 12 public rows | Directional only |
| Coding | 51.6Estimated · #51/152 | 66.5Supported · #12/152 | Directional onlyBenchAlign lane · 2 vs 10 public rows | Directional only |
| Reasoning | Not ranked | 55.5#18/20 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Knowledge | Not ranked | 71.4Supported · #9/183 | Not comparableBenchAlign lane · 0 vs 6 public rows | Not comparable |
| Math | Not ranked | 65.4#2/7 | Not comparableProvisional lane · 0 vs 3 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | 87.8#3/48 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Instruction following | Not ranked | 75.4#61/123 | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
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.
SWE-bench Pro
Coding
Atria Dawn Preview: 59.6%Claude Opus 4.8: 69.2%Normalized gap 9.6Atria Dawn Preview source Claude Opus 4.8 sourceBrowseComp
Agentic
Atria Dawn Preview: 92.5%Claude Opus 4.8: 84.3%Normalized gap 8.2Atria Dawn Preview source Claude Opus 4.8 source
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
- Atria Dawn Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Claude Opus 4.8
- $0.0175
- Fits in one request
Atria Dawn Preview has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Atria Dawn Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Claude Opus 4.8
- $0.325
- Fits in one request
Atria Dawn Preview has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Atria Dawn Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
- Claude Opus 4.8
- $1.35
- Fits in one request
- Cached input priced at the published list-input rate
Claude Opus 4.8 has no published cached-input rate, so cached tokens use its listed input rate. Atria Dawn Preview has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationAtria Dawn PreviewClaude Opus 4.8
Context window
Maximum documented context; output-token limits may be lower.
Atria Dawn Preview
Claude Opus 4.8
API model ID
Atria Dawn Preview
Atria-Dawn-Preview
Atria Dawn Preview model cardClaude Opus 4.8
claude-opus-4-8
Anthropic model overviewCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Atria Dawn Preview
No comparable hosted API rate
Atria Dawn Preview model cardClaude Opus 4.8
Not published
Documented inputs
Atria Dawn Preview
Not sourced
Claude Opus 4.8
text, image
Anthropic model overviewDocumented outputs
Atria Dawn Preview
Not sourced
Claude Opus 4.8
Provider availability
Atria Dawn Preview
Not sourced
Claude Opus 4.8
Generally Available · Claude API
Anthropic model overviewReasoning profile
Atria Dawn Preview
Reasoning
Claude Opus 4.8
Reasoning
Weight access
Atria Dawn Preview
Open Weight
Claude Opus 4.8
Proprietary
License
Atria Dawn Preview
Open Weight
Claude Opus 4.8
Proprietary
Release date
Atria Dawn Preview
2026-09-14
Claude Opus 4.8
2026-05-28
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
- At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Claude Opus 4.8 has the larger documented window (1M).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence48 rows
Agentic
AutomationBench
Atria Dawn Preview53.8%SourceClaude Opus 4.8—Not directly comparable
BFCL v4
Atria Dawn Preview77.0%SourceClaude Opus 4.8—Not directly comparable
CyberGym
Atria Dawn Preview86.5%SourceClaude Opus 4.8—Not directly comparable
DeepSearchQA
Atria Dawn Preview96.0%SourceClaude Opus 4.893.1%SourceAtria Dawn Preview leads this result
BrowseComp
Atria Dawn Preview92.5%SourceClaude Opus 4.884.3%SourceAtria Dawn Preview leads this result
skillsBench
Atria Dawn Preview66.4%SourceClaude Opus 4.8—Not directly comparable
MLE-Bench Lite
Atria Dawn Preview86.2%SourceClaude Opus 4.8—Not directly comparable
WideResearch
Atria Dawn Preview81.9%SourceClaude Opus 4.8—Not directly comparable
τ³-bench results
Atria Dawn Preview41.2%SourceClaude Opus 4.8—Not directly comparable
Terminal-Bench 2.1
Atria Dawn Preview78.3%SourceClaude Opus 4.8—Not directly comparable
JobBench
Atria Dawn Preview50.3%SourceClaude Opus 4.8—Not directly comparable
Terminal-Bench 3.0
Atria Dawn Preview—Claude Opus 4.821.1%SourceNot directly comparable
Terminal-Bench 2.0
Atria Dawn Preview—Claude Opus 4.874.6%SourceNot directly comparable
OSWorld-Verified
Atria Dawn Preview—Claude Opus 4.883.4%SourceNot directly comparable
Finance Agent v2
Atria Dawn Preview—Claude Opus 4.853.9%SourceNot directly comparable
MCP Atlas
Atria Dawn Preview—Claude Opus 4.882.2%SourceNot directly comparable
Toolathlon
Atria Dawn Preview—Claude Opus 4.859.9%SourceNot directly comparable
Gert Labs
Atria Dawn Preview—Claude Opus 4.872.97%SourceNot directly comparable
ResearchClawBench
Atria Dawn Preview—Claude Opus 4.821.1%SourceNot directly comparable
OSWorld 2.0
Atria Dawn Preview—Claude Opus 4.820.6%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Atria Dawn Preview—Claude Opus 4.871.9%SourceNot directly comparable
Coding
Terminal-Bench 2.1
Atria Dawn Preview78.3%SourceClaude Opus 4.8—Not directly comparable
SWE-bench Pro
Atria Dawn Preview59.6%SourceClaude Opus 4.869.2%SourceClaude Opus 4.8 leads this result
SWE-bench Verified
Atria Dawn Preview—Claude Opus 4.888.6%SourceNot directly comparable
SWE Multilingual
Atria Dawn Preview—Claude Opus 4.884.4%SourceNot directly comparable
SWE Multimodal
Atria Dawn Preview—Claude Opus 4.838.4%SourceNot directly comparable
Terminal-Bench 2.0
Atria Dawn Preview—Claude Opus 4.874.6%SourceNot directly comparable
cursorBench31
Atria Dawn Preview—Claude Opus 4.858.4%SourceNot directly comparable
cursorBench32
Atria Dawn Preview—Claude Opus 4.862.3%SourceNot directly comparable
FrontierCode 1.1 Main
Atria Dawn Preview—Claude Opus 4.846.5%SourceNot directly comparable
LiveCodeBench (Vals)
Atria Dawn Preview—Claude Opus 4.887.8%SourceNot directly comparable
SWE-bench (Vals)
Atria Dawn Preview—Claude Opus 4.888.6%SourceNot directly comparable
Reasoning
Knowledge
GPQA
Atria Dawn Preview—Claude Opus 4.893.6%SourceNot directly comparable
GPQA-D
Atria Dawn Preview—Claude Opus 4.893.6%SourceNot directly comparable
HLE
Atria Dawn Preview—Claude Opus 4.857.9%SourceNot directly comparable
HLE w/o tools
Atria Dawn Preview—Claude Opus 4.849.8%SourceNot directly comparable
GPQA Diamond (Vals)
Atria Dawn Preview—Claude Opus 4.892.4%SourceNot directly comparable
MMLU-Pro (Vals)
Atria Dawn Preview—Claude Opus 4.889.6%SourceNot directly comparable
Math
Multilingual
INCLUDE
Atria Dawn Preview—Claude Opus 4.887.6%SourceNot directly comparable
Multimodal
OfficeQA Pro
Atria Dawn Preview—Claude Opus 4.866.2%SourceNot directly comparable
ScreenSpot Pro
Atria Dawn Preview—Claude Opus 4.887.9%SourceNot directly comparable
CharXiv
Atria Dawn Preview—Claude Opus 4.889.9%SourceNot directly comparable
CharXiv w/o tools
Atria Dawn Preview—Claude Opus 4.880.5%SourceNot directly comparable
Frequently asked questions
Which is better, Atria Dawn Preview or Claude Opus 4.8?
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, Atria Dawn Preview or Claude Opus 4.8?
Claude Opus 4.8 scores higher for coding on the public lane, 66.5 to 51.6. Atria Dawn Preview is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, Atria Dawn Preview or Claude Opus 4.8?
Claude Opus 4.8 scores higher for agentic tasks on the public lane, 62.5 to 57.9. Atria Dawn Preview is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which costs less, Atria Dawn Preview or Claude Opus 4.8?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, Atria Dawn Preview or Claude Opus 4.8?
Claude Opus 4.8 has the larger documented context window: 1M, compared with 256K.
Related comparisons
Last updated September 14, 2026
Watch Atria Dawn Preview vs Claude Opus 4.8
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.