Granite 4.2 3B
- Input
- $0.03/M tok
- Output
- $0.12/M tok
- Blended
- $0.052/M tok · 3:1
- Speed
- 220tok/s
- Context
- 131Ktokens
- Cost per task
- ~$0.05/taskest.
Long-tail import from Artificial Analysis (2026-09-24); AA variant: Granite 4.2 3B. Price = AA median host price.
Rankings
Self-hostingEstimated
How we estimate →What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.
- Hardware
- 1× B200 180GB
- BF16 weights
- Throughput
- ~16,546 tok/s
- many requests batched
- Price
- $0.045–0.40 /M tok
- busy → light use
- On your own machine
- RTX 5090 32GB (int4, 300+ tok/s single-stream)
- single consumer GPU or Mac
Assumptions (5)
- BF16 weights (3.66B params) + 25% KV-cache headroom ≈ 9 GB
- 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
- ~16,546 output tok/s aggregate at batch 256 (bandwidth-bound)
- Blended 3:1 input:output; prefill ~145,236 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
~#19 of 21 (estimated)~61.6%
- #1Claude Opus 596.0%
- #2Claude Sonnet 585.2%
- #18gpt-oss-120b62.4%
- ~#19Granite 4.2 3BEstimated61.6%
- #19NVIDIA Nemotron-Labs Dory1.0 4B59.5%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#23 of 24 (estimated)~45.4%
- #1Claude Fable 5.293.0%
- #2Claude Opus 5.589.9%
- #22Muse Glimmer51.2%
- ~#23Granite 4.2 3BEstimated45.4%
- #23Nex-N2.5-Mini43.8%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#43 of 46 (estimated)~32.7%
- #1Gemini 4 Argon78.8%
- #2DeepSeek V4.1 Flash 101777.0%
- #42Nex-N2.5-Mini36.1%
- ~#43Granite 4.2 3BEstimated32.7%
- #43Hy328.0%
No published score — estimated from related benchmarks
Terminal-Bench 4.0Verified
#63 of 1010.0%
- #1Kimi K3.1 Pro67.2%
- #2Claude Sonnet 5.563.6%
- #63Granite 4.2 30B0.0%
- #63Granite 4.2 3B0.0%
- #63Granite 4.2 8B0.0%
artificialanalysis.ai · 2026-09-24
LiveCodeBenchEstimated
~#22 of 28 (estimated)~44.7%
- #1Qwen3.8-Omni-Flash92.6%
- #2Qwen3.8 Flash-Next91.9%
- #21Devstral 244.8%
- ~#22Granite 4.2 3BEstimated44.7%
- #22Llama 4 Maverick39.7%
No published score — estimated from related benchmarks
SciCodeVerified
#108 of 11825.3%
- #1Claude Fable 5.267.7%
- #2Claude Opus 5.566.9%
- #107MiniCPM5-2B26.3%
- #108Granite 4.2 3B25.3%
- #109Ling 3.0 Tiny24.2%
artificialanalysis.ai · 2026-09-24
AA Coding Agent IndexEstimated
~#25 of 26 (estimated)~27.9
- #1Claude Opus 5.561.7
- #2Claude Opus 559.7
- #24GPT-6 Luna31.1
- ~#25Granite 4.2 3BEstimated27.9
- #25Muse Spark Open26.3
No published score — estimated from related benchmarks
Hard reasoning
GPQA DiamondVerified
#144 of 19255.9%
- #1GPT-6 Astra96.1%
- #2Gemini 3.8 Flash95.3%
- #143GLM-4.6V56.6%
- #144Granite 4.2 3B55.9%
- #145LFM2.5-2.6B55.8%
artificialanalysis.ai · 2026-09-24
Humanity's Last ExamVerified
#137 of 1986.6%
- #1Claude Fable 5.274.3%
- #2Claude Opus 5.561.4%
- #136LFM2.5-1.2B-Instruct6.7%
- #137Granite 4.2 3B6.6%
- #138MiniCPM5-1B6.5%
artificialanalysis.ai · 2026-09-24
AIME (latest)Estimated
~#23 of 34 (estimated)~43.7%
- #1GLM-5.3 NVFP499.5%
- #2Beam 1.0 501B A23B97.7%
- #22Phi-4-reasoning-plus78.0%
- ~#23Granite 4.2 3BEstimated43.7%
- #23Mistral Large 338.0%
No published score — estimated from related benchmarks
Agents
τ²-bench (Telecom)Estimated
~#59 of 120 (estimated)~34.2%
- #1Step 3.7 Flash98.5%
- #2DeepSeek-V4-Pro-0813-nvfp4-DSpark98.3%
- #58Sarvam 30B34.5%
- ~#59Granite 4.2 3BEstimated34.2%
- #59Gemma 4 12B31.9%
No published score — estimated from related benchmarks
OSWorld-VerifiedEstimated
~#12 of 15 (estimated)~58.1%
- #1Kimi K384.8%
- #2Qwen3.8 27B84.3%
- #11Muse Glimmer65.9%
- ~#12Granite 4.2 3BEstimated58.1%
- #12Nemotron 3.5 Super VL NVFP452.3%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~last of 5 (estimated)below measured range (<47.5%)
No published score — estimated from related benchmarks
BrowseCompEstimated
~#19 of 22 (estimated)~52.4%
- #1Atria Dawn Preview92.5%
- #2GPT-6 Astra91.5%
- #18Nemotron 3.5 Bolt 74B-A9B70.0%
- ~#19Granite 4.2 3BEstimated52.4%
- #19Mistral Medium 3.548.6%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Verified
#89 of 91188 Elo
- #1Claude Opus 5.51866
- #2Muse Wonder Full1857
- #88Ministral 3 8B200
- #89Granite 4.2 3B188
- #90K2 Think V2117
artificialanalysis.ai · 2026-09-24