Claude 3.5 Haiku vs Magistral Small 1.2
Wins 3 of 7 areas
Agents · Facts · Long documents
Wins 4 of 7 areas
Coding · Reasoning · Images and charts · Following instructions
Magistral Small 1.2 wins more areas, narrowly.Claude 3.5 Haiku is better at agents.
Scores updated · 18 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- ReasoningHard problems that need careful thinking02Magistral Small 1.22 of 3 tests · 1 tie
- CodingWriting and fixing software12Magistral Small 1.22 of 3 tests
- Images and chartsUnderstanding pictures, charts and video01Magistral Small 1.21 of 1 test
- Following instructionsDoing exactly what it is asked01Magistral Small 1.21 of 1 test
- AgentsCarrying out multi-step tasks on its own10Claude 3.5 Haiku1 of 2 tests · 1 tie
- FactsGetting facts right instead of making them up10Claude 3.5 Haiku1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10Claude 3.5 Haiku1 of 1 test
Images and charts, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Magistral Small 1.2 pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+25.5points ahead
- Harder college exam questions with imagesMMMU-Pro+9.9points ahead
- Code for real scientific research problemsSciCode+7.8points ahead
Where Claude 3.5 Haiku pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+49.7points ahead
- Reasons across sets of long documentsAA-LCR+8points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+5.6points ahead
Every test, side by side
All 18 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingMagistral Small 1.2
- SciCodeMagistral Small 1.2 by 7.827.435.2+7.8
- Terminal-Bench 2.1Claude 3.5 Haiku by 5.610.14.5+5.6
- Terminal-Bench HardMagistral Small 1.2 by 2.22.34.5+2.2
AgentsClaude 3.5 Haiku
- τ-Bench V3 · BankingClaude 3.5 Haiku by 1.35.84.5+1.3
- GDPValtie00tie
ReasoningMagistral Small 1.2
- GPQA DiamondMagistral Small 1.2 by 25.540.866.3+25.5
- Humanity's Last ExamMagistral Small 1.2 by 2.83.66.4+2.8
- CritPttie00.3tie
FactsClaude 3.5 Haiku
- AA-Omniscience · Non-hallucinationClaude 3.5 Haiku by 49.758.99.2+49.7
- AA-Omniscience · Accuracytie13.213.4tie
Images and chartsMagistral Small 1.2
- MMMU-ProMagistral Small 1.2 by 9.945.655.5+9.9
Following instructionsMagistral Small 1.2
- IFBenchMagistral Small 1.2 by 1.642.844.4+1.6
Other results5 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- AA-OmniscienceClaude 3.5 Haiku by 42.6-22.5-65.1+42.6
- τ²-Bench Telecom (AA run)Magistral Small 1.2 by 3.224.627.8+3.2
- Artificial Analysis Coding IndexClaude 3.5 Haiku by 1.215.914.7+1.2
- AA Agentic Indextie1.91.5tie
- AA Intelligencetie8.98.6tie
Questions people ask
Which is better, Claude 3.5 Haiku or Magistral Small 1.2?
Magistral Small 1.2 wins four of the seven areas where both have results: coding, reasoning, images and charts and following instructions. Claude 3.5 Haiku wins agents, facts and long documents.
Which is better for coding?
Magistral Small 1.2. It wins 2 of the 3 coding tests both models report; Claude 3.5 Haiku wins 1.
How do you compare the two?
We use the 18 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 5 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.