Skip to content
Model Pareto

Bonsai 27B

PrismMLOpen weights1-bit (Q1_0_g128)huggingface.co ↗
Compare with nearest rivalsCompare with…
Input
—
Output
—
Speed
~84tok/sest.
Context
262Ktokens

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
1× B200 180GB
1-bit (Q1_0_g128) weights
Throughput
~6,454 tok/s
many requests batched
Price
$0.13–1.15 /M tok
busy → light use
On your own machine
RTX 5090 32GB (1-bit (Q1_0_g128), 300+ tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • 1-BIT (Q1_0_G128) weights (27.3B params) + 25% KV-cache headroom ≈ 5 GB
  • 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~6,454 output tok/s aggregate at batch 184 (bandwidth-bound)
  • Blended 3:1 input:output; prefill ~38,942 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

SWE-bench VerifiedEstimated
~#5 of 14 (estimated)~79.4%
  1. #1Claude Opus 596.0%
  2. #2Claude Sonnet 585.2%
  3. #4MiniMax-M380.5%
  4. ~#5Bonsai 27BEstimated79.4%
  5. #5Hy378.0%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#9 of 19 (estimated)~63.2%
  1. #1Claude Opus 5.589.9%
  2. #2Claude Sonnet 5.581.3%
  3. #8Claude Sonnet 563.2%
  4. ~#9Bonsai 27BEstimated63.2%
  5. #9Nex-N2.5-Pro61.2%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#29 of 32 (estimated)~47.9%
  1. #1GPT-6.1 Sol75.2%
  2. #2DeepSeek V4.1 Flash74.2%
  3. #28Qwen3.8 Max (0902)51.0%
  4. ~#29Bonsai 27BEstimated47.9%
  5. #29Qwen3.8 27B42.2%
No published score — estimated from related benchmarks
Terminal-Bench 4.0Estimated
~#19 of 85 (estimated)~21.0%
  1. #1Claude Sonnet 5.563.6%
  2. #2Claude Opus 5.559.6%
  3. #18MiMo-V2.6-Flash22.7%
  4. ~#19Bonsai 27BEstimated21.0%
  5. #19Gemini 3.8 Flash19.7%
No published score — estimated from related benchmarks
SciCodeEstimated
~#34 of 103 (estimated)~49.5%
  1. #1Claude Opus 5.566.9%
  2. #2Claude Fable 5.163.1%
  3. #33Inkling Small49.7%
  4. ~#34Bonsai 27BEstimated49.5%
  5. #34Hy348.6%
No published score — estimated from related benchmarks
AA Coding Agent IndexEstimated
~#10 of 15 (estimated)~50.3
  1. #1Claude Opus 5.566.0
  2. #2Claude Fable 5.162.2
  3. #9Kimi K351.9
  4. ~#10Bonsai 27BEstimated50.3
  5. #10Grok 4.647.0
No published score — estimated from related benchmarks

Writing & chat

MMMLUEstimated
~#3 of 5 (estimated)~85.1%
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. ~#3Bonsai 27BEstimated85.1%
  4. #3Claude Haiku 4.583.0%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#33 of 54 (estimated)~1415 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #32Muse GlimmerFew votes1418
  4. ~#33Bonsai 27BEstimated1415
  5. #33MiMo-V2.51411
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#33 of 54 (estimated)~1429 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #32MiMo-V2.51434
  4. ~#33Bonsai 27BEstimated1429
  5. #33Mistral Medium 3.51426
No published score — estimated from related benchmarks
EQ-Bench Creative Writing v3 (Elo)Estimated
~#10 of 26 (estimated)~1895 Elo
  1. #1GPT-6 Astra2173
  2. #2Claude Fable 5.12162
  3. #9Muse Spark 1.31906
  4. ~#10Bonsai 27BEstimated1895
  5. #10GPT-5.6 Terra1855
No published score — estimated from related benchmarks

Vision

MMMUEstimated
~#2 of 3 (estimated)~71.3%
  1. #1Claude Haiku 4.573.2%
  2. ~#2Bonsai 27BEstimated71.3%
  3. #2Llama 4 Maverick69.4%
No published score — estimated from related benchmarks
CharXiv ReasoningEstimated
~#8 of 9 (estimated)~75.9%
  1. #1Gemini 3.8 Flash86.2%
  2. #2Kimi K384.8%
  3. #7Claude Sonnet 577.0%
  4. ~#8Bonsai 27BEstimated75.9%
  5. #8Gemini 3.5 Flash-Lite74.5%
No published score — estimated from related benchmarks
OmniDocBench v1.5Estimated
~#5 of 8 (estimated)~82.4%
  1. #1MiniMax-M391.6%
  2. #2Kimi K391.1%
  3. #4Gemini 3.1 Pro85.3%
  4. ~#5Bonsai 27BEstimated82.4%
  5. #5Claude Haiku 4.579.6%
No published score — estimated from related benchmarks
GDP.pdf (AA)Estimated
~#30 of 43 (estimated)~9.3%
  1. #1GPT-6 Astra31.0%
  2. #2Muse Spark 1.326.6%
  3. #29MiniMax-M39.8%
  4. ~#30Bonsai 27BEstimated9.3%
  5. #30Ling-3.0-flash-VL8.4%
No published score — estimated from related benchmarks
ChartographyEstimated
~#20 of 23 (estimated)~11.6%
  1. #1GPT-6 Astra71.0%
  2. #2Claude Opus 5.566.3%
  3. #19DeepSeek V4.1 Flash11.6%
  4. ~#20Bonsai 27BEstimated11.6%
  5. #20Gemini 3.5 Flash-Lite10.2%
No published score — estimated from related benchmarks

Hard reasoning

GPQA DiamondEstimated
~#70 of 180 (estimated)~79.9%
  1. #1GPT-6 Astra96.1%
  2. #2Gemini 3.8 Flash95.3%
  3. #69Nemotron 3 Super 120B A12B80.0%
  4. ~#70Bonsai 27BEstimated79.9%
  5. #70EXAONE 4.5 33B79.4%
No published score — estimated from related benchmarks
Humanity's Last ExamEstimated
~#53 of 182 (estimated)~25.2%
  1. #1Claude Opus 5.561.4%
  2. #2Claude Fable 5.159.1%
  3. #52Qwen3.5 122B A10B25.2%
  4. ~#53Bonsai 27BEstimated25.2%
  5. #53Ling 3.0 Flash23.7%
No published score — estimated from related benchmarks
AA Intelligence IndexEstimated
~#30 of 78 (estimated)~29.0
  1. #1Claude Opus 5.557.6
  2. #2Claude Sonnet 5.556.0
  3. #29MiniMax-M329.2
  4. ~#30Bonsai 27BEstimated29.0
  5. #30Quasar 438B26.7
No published score — estimated from related benchmarks