Agentic
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 43.2
- Estimated · #97/152
- Basis
- BenchAlign lane · 0 vs 15 public rows
- Reading
- Not comparable
1-bit Bonsai 27B vs Claude Opus 4.5
Updated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. Use the documented cost, context, and runtime rows instead.
0 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
1-bit Bonsai 27B
1-bit Bonsai 27B 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
1-bit Bonsai 27B is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
1-bit Bonsai 27B is not ranked on the public lane for agentic, so no winner is named for agentic.
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
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Claude Opus 4.5 does not fit this workload in one request. Claude Opus 4.5 has no published cached-input rate, so cached tokens use its listed input rate. 1-bit Bonsai 27B has no comparable published API token rate.
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.
45 Claude Opus 4.5 only
- Shared results
- 0
- 1-bit Bonsai 27B only
- 0
- Claude Opus 4.5 only
- 45
- Like-for-like categories
- 0 / 8
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
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 56.6
- Estimated · #36/151
- Basis
- BenchAlign lane · 0 vs 5 public rows
- Reading
- Not comparable
Reasoning
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 72.1
- Unranked · 4 rankable rows
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 53.9
- Estimated · #61/183
- Basis
- BenchAlign lane · 0 vs 6 public rows
- Reading
- Not comparable
Math
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 58.3
- #6/7
- Basis
- Provisional lane · 0 vs 4 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 82.9
- #2/12
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 23.5
- #46/48
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- 1-bit Bonsai 27B
- Not ranked
- Claude Opus 4.5
- 39.5
- #105/123
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
| Category | 1-bit Bonsai 27B | Claude Opus 4.5 | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | 43.2Estimated · #97/152 | Not comparableBenchAlign lane · 0 vs 15 public rows | Not comparable |
| Coding | Not ranked | 56.6Estimated · #36/151 | Not comparableBenchAlign lane · 0 vs 5 public rows | Not comparable |
| Reasoning | Not ranked | 72.1Unranked · 4 rankable rows | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Knowledge | Not ranked | 53.9Estimated · #61/183 | Not comparableBenchAlign lane · 0 vs 6 public rows | Not comparable |
| Math | Not ranked | 58.3#6/7 | Not comparableProvisional lane · 0 vs 4 weighted rows | Not comparable |
| Multilingual | Not ranked | 82.9#2/12 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Multimodal | Not ranked | 23.5#46/48 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Instruction following | Not ranked | 39.5#105/123 | Not comparableProvisional lane · 0 vs 1 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.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
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
- 1-bit Bonsai 27B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Claude Opus 4.5
- $0.0175
- Fits in one request
1-bit Bonsai 27B has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- 1-bit Bonsai 27B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Claude Opus 4.5
- $0.325
- Fits in one request
1-bit Bonsai 27B has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- 1-bit Bonsai 27B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
- Claude Opus 4.5
- $1.35
- Does not fit in one request
- Cached input priced at the published list-input rate
Claude Opus 4.5 does not fit this workload in one request. Claude Opus 4.5 has no published cached-input rate, so cached tokens use its listed input rate. 1-bit Bonsai 27B 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.
Specification1-bit Bonsai 27BClaude Opus 4.5
Context window
Maximum documented context; output-token limits may be lower.
1-bit Bonsai 27B
Claude Opus 4.5
200K
API model ID
1-bit Bonsai 27B
prism-ml/Bonsai-27B-gguf
PrismML 1-bit Bonsai 27B model cardClaude Opus 4.5
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
1-bit Bonsai 27B
No comparable hosted API rate
PrismML 1-bit Bonsai 27B model cardClaude Opus 4.5
Not published
Documented inputs
1-bit Bonsai 27B
Not sourced
Claude Opus 4.5
Not sourced
Documented outputs
1-bit Bonsai 27B
Not sourced
Claude Opus 4.5
Not sourced
Provider availability
1-bit Bonsai 27B
Not sourced
Claude Opus 4.5
Not sourced
Reasoning profile
1-bit Bonsai 27B
Reasoning
Claude Opus 4.5
Non-Reasoning
Weight access
1-bit Bonsai 27B
Open Weight
Claude Opus 4.5
Proprietary
License
1-bit Bonsai 27B
Open Weight
Claude Opus 4.5
Proprietary
Release date
1-bit Bonsai 27B
2026-07-14
Claude Opus 4.5
2025-11-01
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
- The public evidence has no benchmark result shared by both models, so it does not support a quality verdict.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- 1-bit Bonsai 27B has the larger documented window (262K).
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 evidence45 rows
Agentic
Terminal-Bench 2.0
1-bit Bonsai 27B—Claude Opus 4.559.3%SourceNot directly comparable
OSWorld-Verified
1-bit Bonsai 27B—Claude Opus 4.566.3%SourceNot directly comparable
OSWorld
1-bit Bonsai 27B—Claude Opus 4.566.3%SourceNot directly comparable
Claw-Eval
1-bit Bonsai 27B—Claude Opus 4.559.6%SourceNot directly comparable
QwenClawBench
1-bit Bonsai 27B—Claude Opus 4.552.3%SourceNot directly comparable
τ³-bench results
1-bit Bonsai 27B—Claude Opus 4.570.2%SourceNot directly comparable
VITA-Bench
1-bit Bonsai 27B—Claude Opus 4.523.3%SourceNot directly comparable
DeepPlanning
1-bit Bonsai 27B—Claude Opus 4.526.4%SourceNot directly comparable
Toolathlon
1-bit Bonsai 27B—Claude Opus 4.543.5%SourceNot directly comparable
MCP Atlas
1-bit Bonsai 27B—Claude Opus 4.542.3%SourceNot directly comparable
MCP-Tasks
1-bit Bonsai 27B—Claude Opus 4.571.8%SourceNot directly comparable
WideResearch
1-bit Bonsai 27B—Claude Opus 4.576.4%SourceNot directly comparable
CyberGym
1-bit Bonsai 27B—Claude Opus 4.550.6%SourceNot directly comparable
Gert Labs
1-bit Bonsai 27B—Claude Opus 4.564.23%SourceNot directly comparable
JobBench
1-bit Bonsai 27B—Claude Opus 4.532.3%SourceNot directly comparable
Coding
SWE-bench Verified
1-bit Bonsai 27B—Claude Opus 4.580.9%SourceNot directly comparable
LiveCodeBench v6
1-bit Bonsai 27B—Claude Opus 4.584.8%SourceNot directly comparable
SWE-bench Pro
1-bit Bonsai 27B—Claude Opus 4.557.1%SourceNot directly comparable
SWE Multilingual
1-bit Bonsai 27B—Claude Opus 4.577.5%SourceNot directly comparable
NL2Repo
1-bit Bonsai 27B—Claude Opus 4.543.2%SourceNot directly comparable
Reasoning
Knowledge
GPQA
1-bit Bonsai 27B—Claude Opus 4.587%SourceNot directly comparable
SuperGPQA
1-bit Bonsai 27B—Claude Opus 4.570.6%SourceNot directly comparable
MMLU-Pro
1-bit Bonsai 27B—Claude Opus 4.589.5%SourceNot directly comparable
MMLU-Redux
1-bit Bonsai 27B—Claude Opus 4.596.6%SourceNot directly comparable
C-Eval
1-bit Bonsai 27B—Claude Opus 4.592.2%SourceNot directly comparable
HLE
1-bit Bonsai 27B—Claude Opus 4.530.8%SourceNot directly comparable
Math
AIME26
1-bit Bonsai 27B—Claude Opus 4.595.1%SourceNot directly comparable
HMMT Feb 2025
1-bit Bonsai 27B—Claude Opus 4.592.9%SourceNot directly comparable
HMMT Nov 2025
1-bit Bonsai 27B—Claude Opus 4.593.3%SourceNot directly comparable
HMMT Feb 2026
1-bit Bonsai 27B—Claude Opus 4.585.3%SourceNot directly comparable
MMAnswerBench
1-bit Bonsai 27B—Claude Opus 4.584.0%SourceNot directly comparable
FrontierMath v2 (Tiers 1-3)
1-bit Bonsai 27B—Claude Opus 4.520.690%SourceNot directly comparable
FrontierMath v2 (Tier 4)
1-bit Bonsai 27B—Claude Opus 4.54.167%SourceNot directly comparable
Multilingual
Multimodal
MMMU-Pro
1-bit Bonsai 27B—Claude Opus 4.570.6%SourceNot directly comparable
MathVision
1-bit Bonsai 27B—Claude Opus 4.574.3%SourceNot directly comparable
CharXiv
1-bit Bonsai 27B—Claude Opus 4.568.5%SourceNot directly comparable
VideoMMMU
1-bit Bonsai 27B—Claude Opus 4.584.4%SourceNot directly comparable
ScreenSpot Pro
1-bit Bonsai 27B—Claude Opus 4.545.7%SourceNot directly comparable
V*
1-bit Bonsai 27B—Claude Opus 4.567.0%SourceNot directly comparable
Instruction following
Frequently asked questions
Which is better, 1-bit Bonsai 27B or Claude Opus 4.5?
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, 1-bit Bonsai 27B or Claude Opus 4.5?
1-bit Bonsai 27B is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, 1-bit Bonsai 27B or Claude Opus 4.5?
1-bit Bonsai 27B is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, 1-bit Bonsai 27B or Claude Opus 4.5?
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, 1-bit Bonsai 27B or Claude Opus 4.5?
1-bit Bonsai 27B has the larger documented context window: 262K, compared with 200K.
Related comparisons
Last updated September 10, 2026
Watch 1-bit Bonsai 27B vs Claude Opus 4.5
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.